Why States Unite Their Efforts Against Platforms

How can one state influence a platform used nationwide? This course explains how U.S. federalism enables state attorneys general (AGs) to coordinate investigations and lawsuits when federal action falls behind technological change. Its central thesis is qualified: coalitions can pool resources and negotiate reforms with nationwide practical effects, but they do not automatically create national law or prove wrongdoing. You will map the roles of states, Congress, agencies, courts, and private plaintiffs; connect harms to consumer-protection and privacy authorities; follow the eight-stage path from allegation to compliance; and apply it to the 2023 Meta action and historical examples. You will also evaluate evidence, settlements, preemption, Section 230, the First Amendment, causation, and democratic accountability.

Chapter 1: Federalism Under Pressure—Who Has the Power to Act?

1.1: Shared Sovereignty, Different Jobs—Reading the Federalism Map

Jordan Ellis begins with a reasonable intuition: if a platform operates across the country, perhaps only Congress can regulate it. The federalism map tests that intuition—but also limits it. A national platform may face action from states, federal institutions, courts, and private plaintiffs at the same time. Yet state action remains state action. It does not automatically become a national statute, and state officials do not possess unlimited authority simply because a company operates nationwide.

%FULL[infograph] %A clean, text-supported diagram showing the U.S. Constitution at the center, federal institutions on one side, state institutions on the other, and overlapping areas for consumer protection and antitrust enforcement.

Learning objectives: By the end of this subchapter, you should be able to:

• Define federalism, concurrent authority, divided authority, and state police powers.
• Distinguish the roles of Congress, federal agencies, courts, state legislatures, State attorneys general (AGs), and private plaintiffs.
• Use an actor-authority-outcome map to explain how state action can affect a national platform without becoming federal legislation.

Federalism as an Allocation of Responsibility

Federalism is a system in which power is shared between the national government and individual state governments. The important word is shared. Federalism is not a simple wall dividing federal subjects from state subjects. The two levels may overlap, cooperate, compete, or disagree. This overlapping structure is a basic feature of American governance, not necessarily a legal error. #(Elazar, 1987, Exploring Federalism)

A useful distinction is between formal legal authority and practical spillover. Formal legal authority is the power an institution is legally permitted to exercise. Practical spillover is an effect that reaches beyond the formal boundaries of the action. For example, a state may enforce its own consumer-protection law against a platform. If the platform uses one national design, it may change that design for users everywhere. The change may have national practical effects, but the state has not enacted federal legislation.

Three related terms help organize the map:

Concept	Plain-language meaning	Important boundary
Federalism	Power shared between the national government and the states	A state concern does not automatically create authority over every related activity
Concurrent authority	Federal and state governments may both act in an area	The governments need not act together, and conflicts may arise
Divided authority	Different institutions hold different parts of the governing process	No single institution normally performs all legislative, investigative, adjudicative, and implementation functions
State police powers	State authority associated with health, safety, and welfare	Police powers do not override valid federal law or other constitutional limits

Concurrent authority is therefore an overlap in legal power. Consumer protection and antitrust enforcement are examples of areas in which state and federal governments may both act. Concurrent authority does not mean that federal and state officials must coordinate. They may investigate separately, bring separate actions, or adopt different enforcement priorities. It also does not eliminate federal preemption, the possibility that valid federal law displaces or limits conflicting state law or state action. Overlapping authority can produce cooperation, but it can also produce disputes about preemption, jurisdiction, and remedies. Here, jurisdiction means an institution’s legal power to hear a matter or take a particular action. #(Bednar, 2009, The Robust Federation)

State police powers are state powers associated with protecting health, safety, and welfare. They are connected to powers retained by the states under the Tenth Amendment. In practical terms, these powers help explain why a state may address public harms even when the affected company operates across state lines. But the Tenth Amendment is not an independent grant of unlimited litigation power to every State attorney general. Nor do police powers automatically prevail over valid federal law.

Other constitutional limits may matter as well. The Dormant Commerce Clause refers to a constitutional principle that can restrict certain state measures affecting interstate commerce, including measures that discriminate against interstate commerce or impose constitutionally impermissible burdens on it. Whether a particular state action violates that principle depends on the facts, the legal theory, and the relevant law. The basic lesson is straightforward: a state’s important policy goal and its legal authority are related, but they are not the same question.

Federal delay can create a regulatory gap—a period in which existing rules do not fully address a developing problem or in which federal legislation or enforcement has not yet supplied a comprehensive response. A regulatory gap does not eliminate legal constraints. States may use authority already available under state law, but they cannot treat the gap as permission to regulate everything, disregard preemption, or bypass constitutional limits.

From Institutional Authority to Practical Effects

Federalism identifies the relationship between levels of government. Divided authority adds a second layer: different institutions at those levels have different jobs. Congress creates prospective rules. Agencies implement and enforce statutes. Courts adjudicate disputes. State legislatures create state law. State AGs represent state interests and enforce laws protecting the public. Private plaintiffs pursue individual or class remedies. These roles overlap in a broader system, but they are not interchangeable. #(Vile, 1967, Constitutionalism and the Separation of Powers)

The following actor-authority-outcome map connects each institution to its usual function:

Actor	Formal authority and tool	Typical outcome	Key limitation
Congress	Creates prospective federal rules through legislation	A national statutory requirement	Must act through the legislative process and within constitutional authority
Federal agency	Implements and enforces statutes; may issue rules where Congress has delegated authority	Administrative rules, investigations, or enforcement actions	Cannot exceed its statutory authority
Court	Adjudicates disputes and may issue judgments or approve a consent decree	A judicial ruling, injunction, or enforceable court order	Decides the case before it; does not function as a legislature
State legislature	Creates state statutes, including UDAP laws or State privacy laws	State-wide legal duties within that state	Law remains subject to federal and constitutional limits
State AG	Advises state government and enforces public-protection laws; may investigate or litigate where state law permits	Investigation, litigation, settlement, or negotiated compliance	Powers differ by state and do not make the AG a legislature or agency
Private plaintiff	Brings an individual or class action seeking a private remedy	Compensation, injunctive relief, or another court-ordered remedy when available	Must satisfy applicable procedural and legal requirements

UDAP laws are laws addressing unfair or deceptive acts or practices. A State AG may enforce such a law only when the state’s law gives that office the relevant authority. Depending on state law, an AG may use a civil investigative demand (CID)—a formal request for information—or a subpoena. A multistate investigation is an investigation involving officials from more than one state. A multistate coalition is the participating group. The coalition can coordinate strategy, but each AG remains bounded by that state’s law.

Consider the possible path when several states are concerned about a platform’s treatment of young users:

flowchart TD
  n1["Shared concern about platform conduct"]
  n2["Authority check: state law and Constitution"]
  n3["Separate or coalition action by State AGs"]
  n4["Investigation: CID or subpoena if authorized"]
  n5["Litigation: court adjudicates"]
  n6["Negotiation: settlement or assurance"]
  n7["Judgment or court-ordered remedy"]
  n8["Negotiated compliance obligation"]
  n9["Practical spillover: platform may change nationwide"]
  n10["Action narrowed or blocked by legal limits"]
  n1 --> n2
  n2 --> n3
  n2 --> n10
  n3 --> n4
  n4 --> n5
  n4 --> n6
  n5 --> n7
  n6 --> n8
  n7 --> n9
  n8 --> n9

This pathway separates legal stages that are often confused. A complaint contains allegations, not automatically proven facts. Evidence is information used to support or contest those allegations; in litigation, discovery is the court-supervised exchange of information relevant to the dispute. An adjudicated finding is a conclusion reached by a court through a legal proceeding.

A settlement is a negotiated resolution. A consent decree is a negotiated resolution entered or approved by a court. An assurance of voluntary compliance is another form of negotiated commitment whose legal effect depends on applicable law. A resulting compliance obligation may require specified changes, but the obligation is defined by the settlement, decree, or governing law. Negotiated terms do not automatically establish that every allegation was proven or create a nationwide legal precedent.

The practical effects can nevertheless be substantial. A national platform may decide that operating one system is easier than maintaining different systems for different states. It may therefore adopt a nationwide change after facing state investigations or settlements. That result can look like a de facto national policy: a policy that operates nationally in practice without being enacted by Congress. It is still practical spillover, not federal legislation. The platform might instead create state-specific changes, challenge the action in court, or negotiate different commitments. The outcome is therefore possible, not guaranteed.

This distinction also clarifies the State AG’s role. An AG is the chief legal officer of a state who advises state government and enforces laws protecting the public. An AG may investigate, file litigation, negotiate a settlement, or represent the state’s interests. The AG cannot simply create a new statute, perform every function of a regulatory agency, or decide the dispute as a court. State AG powers are not identical across the country; the relevant state constitution, statutes, and procedures matter. Research on State attorneys general emphasizes that they can participate in national policymaking through these bounded state institutions, rather than by replacing Congress or federal agencies. #(Nolette, 2015, Federalism on Trial: State Attorneys General and National Policymaking in Contemporary America)

A brief formative check applies the map to a proposed national rule for youth privacy. Congress would be the institution positioned to create the prospective national rule. A federal agency could implement or enforce that rule, and issue regulations only if authorized. A court would adjudicate a challenge or enforce an available judicial remedy. A state legislature could enact a separate state privacy law. A State AG could enforce an existing state law or investigate and litigate where authorized. A private plaintiff could pursue an individual or class remedy when the law provides one. Assigning all of these tasks to an AG—or treating a state lawsuit as a new national statute—would confuse institutional roles.

Conclusion

Federalism allows national and state governments to act in the same policy area, but shared sovereignty is not unlimited sovereignty. Concurrent authority explains how consumer protection and antitrust enforcement may involve both levels of government. State police powers explain why states can address health, safety, and welfare concerns, while preemption, the Dormant Commerce Clause, the Tenth Amendment’s limited role, and other legal rules preserve boundaries.

The actor-authority-outcome map supplies a method for reading later enforcement disputes: identify who is acting, locate that actor’s formal authority, and then separate the legal outcome from any practical spillover. Congress legislates, agencies administer, courts adjudicate, state legislatures make state law, State AGs represent state interests and enforce authorized laws, and private plaintiffs seek private remedies. A coalition may create coordinated pressure on a national platform, and that pressure may produce nationwide changes, but it does not become federal legislation merely because its effects cross state lines. These distinctions provide the institutional vocabulary for the Chapter 2 enforcement map and for the later eight-stage enforcement pathway.

1.2: Turning Public Concern into a Legal Claim

The federalism map shows that states and the national government can participate in technology governance at the same time, but participation depends on a usable legal hook. Maya Rivera begins with a headline-level concern such as “a platform harms teenagers” and separates it into questions about deception, unfairness, data collection, causation, and legal remedy. A public concern can justify investigation or legislation, but it is not automatically a cause of action. A legally cognizable claim is a legal theory recognized by the applicable law, supported by facts that satisfy its required elements, brought by an authorized actor, and connected to an available remedy.

Learning objectives: By the end of this subchapter, you should be able to:

• Explain how broad concerns must be translated into specific legal claims.
• Differentiate UDAP laws, state privacy laws, common-law claims, federally authorized state enforcement, and parens patriae.
• Recognize why statutory scope, standing, evidence, cause of action, and remedy matter more than a general appeal to public welfare.

The Legal Bases of State Enforcement

Federalism is the allocation of governmental authority between the national government and the states. A state’s police powers are its general powers to protect residents’ health, safety, and welfare. Concurrent authority means that state and federal institutions may act in the same general policy area, but through different legal sources and subject to different limits. Neither concept means that every government may regulate every harmful practice.

For a state attorney general (AG), the first question is not simply whether a practice appears harmful. It is whether the AG has formal legal authority—authority recognized by a statute, common-law doctrine, or federal law—to investigate, litigate, or negotiate. The following comparison shows why the five legal bases should not be treated as interchangeable.

Legal basis	Authority and actor	Main target	Evidence and principal limits	Typical tool or outcome
UDAP laws	State AG under a state unfair or deceptive acts and practices statute	A specific unfair or deceptive business practice	Representations, disclosures, challenged practices, and their connection to the statutory test; social harm alone is insufficient	Investigation, litigation, or statute-specific negotiated compliance
State privacy laws	State AG, agency, private party, or another authorized actor, depending on the statute	Collection, use, disclosure, retention, or control of covered personal information	Information categories, covered entity, conduct, definitions, authorization rules, and exemptions; privacy concern alone is not proof of violation	Enforcement structure and penalties vary by statute
Common-law authority	Courts apply common-law claims; AG authority depends on jurisdiction	Recognized claims such as fraud, negligent misrepresentation, public nuisance, or unjust enrichment	Jurisdiction-specific elements, claim availability, causation, and fit between injury and remedy	Litigation or other relief recognized by the jurisdiction
Federally authorized state enforcement	State AG acts under a federal statute’s express or limited authorization	A specified federal-law violation	The federal statute’s text, conditions, procedures, and limits; authorization is not unrestricted	Civil action or damages where the statute permits
Parens patriae	State acts through the AG to protect certain quasi-sovereign interests	Statewide interests, including residents’ health and welfare	The state interest must be distinct from simply representing every individual injury	Relief tied to the state interest and governing authority

UDAP laws. State unfair or deceptive acts and practices statutes are a foundational route for consumer protection. Many were modeled in part on the Federal Trade Commission Act, but states differ in scope, remedies, definitions, and AG authority. A practice is not automatically “unfair” or “deceptive” merely because it produces an undesirable or socially harmful result. A potential claim ordinarily requires evidence connecting a representation or business practice to the relevant statutory definition. This focus on a concrete information or market problem reflects consumer-protection scholarship. #(Beales, Craswell, and Salop, 1981, The Efficient Regulation of Consumer Information)

State privacy laws. Privacy statutes offer a more targeted route when the alleged harm concerns personal information. State-specific laws, including the California Consumer Privacy Act, may differ in definitions, exemptions, enforcement structures, and penalties. A sound analysis identifies the covered entity, protected information, relevant conduct, statutory duty or prohibition, and any exemption. Privacy regulation operates through varied legal and institutional arrangements rather than one uniform model. #(Bamberger and Mulligan, 2011, Privacy on the Books and on the Ground) A general claim of “privacy invasion” must also be made more precise: unwanted observation, disclosure, information use, and loss of control can describe different legal concerns. #(Solove, 2006, A Taxonomy of Privacy)

Common-law authority. Common law consists of judge-developed legal rules and recognized causes of action. Possible technology-related claims include fraud, negligent misrepresentation, public nuisance, and unjust enrichment. These labels do not carry identical elements. Fraud and negligent misrepresentation may concern false information but can differ in required knowledge or care. Unjust enrichment asks whether a retained benefit is legally unjust. Public nuisance is not a universal technology claim; its availability depends on the jurisdiction, the public interest affected, and the required elements. Tort analysis therefore begins with the legally defined injury and elements, not with the familiarity of the label. #(Dobbs, Hayden, and Bublick, 2011, The Law of Torts)

Federally authorized state enforcement. Sometimes a state AG’s authority comes directly from federal legislation rather than ordinary state police powers. Certain COPPA provisions permit state AG civil actions in specified circumstances. Clayton Act section 4C also provides for parens patriae antitrust actions for damages under stated conditions. These examples show that federal law can create a defined state role, but the federal statute controls the scope of that role. A federal statute touching technology does not automatically give states unrestricted enforcement power.

Parens patriae. This doctrine allows a state to protect certain quasi-sovereign interests, including the health and welfare of its residents. It does not make the AG the general private lawyer for every resident or eliminate the need to identify a legally recognized state interest. Nor does it automatically permit a state to sue the federal government on behalf of its citizens. Standing and state institutional interests remain distinct from generalized public concern. #(Fallon, Manning, Meltzer, and Shapiro, 2015, The Federal Courts and the Federal System)

%HALF left[infograph] %A decision-tree illustration moving from a perceived harm to possible legal bases: UDAP, privacy statute, common law, COPPA-related authority, or antitrust authority, with a reminder that each branch has jurisdiction-specific elements.

Other legal limits may also matter. Federal preemption refers generally to federal law displacing conflicting state law. The Dormant Commerce Clause may constrain certain state measures affecting interstate commerce. Their application is fact-specific and jurisdiction-specific; identifying a possible state-law basis does not resolve those questions.

From Social Harm to a Cognizable Claim

The movement from concern to claim is a narrowing process. The larger eight-stage enforcement pathway eventually reaches investigation, litigation, settlement, and implementation, but this subchapter focuses on the claim-formation stage: matching conduct to authority, evidence, and remedy.

flowchart LR
  n1["Public concern"] --> n2["Specific alleged conduct"]
  n2 --> n3["Legal source and jurisdiction"]
  n3 --> n4["Elements and evidence"]
  n4 --> n5["Standing or enforcement authority"]
  n5 --> n6["Available remedy"]
  n6 --> n7["Formal action or negotiated outcome"]

A practical analysis proceeds in five connected steps:

1. Identify the alleged conduct. Replace “the platform harms teenagers” with a specific allegation about a representation, data practice, disclosure, omission, or other conduct by an identifiable actor.
2. Select the legal source. Ask whether the conduct is best examined under a UDAP statute, privacy statute, common-law claim, federally authorized action, or parens patriae theory. The relevant jurisdiction—the state and court whose law and authority apply—matters at this stage.
3. Match facts to elements. Evidence must address the chosen theory. A UDAP theory may require public representations, user disclosures, the challenged practice, and facts connecting that practice to deception or unfairness. A privacy theory may require the categories of information collected, the platform’s disclosures, the covered entity, statutory definitions, and applicable exemptions. Where required, the analysis must also address causation and the legally recognized injury.
4. Confirm authority and standing. Standing concerns whether the claimant has the legally sufficient connection to invoke a court’s power. Enforcement authority asks whether the AG or another institution is authorized to investigate or sue. Parens patriae may address a state’s quasi-sovereign interest, but it is not a substitute for every individual’s standing.
5. Match the remedy and institution. An AG generally enforces existing law; an AG does not simply create a new legal duty because public concern is intense. A legislature may enact a new statute, an agency may issue rules when authorized, and a court may interpret disputed elements. A settlement may create compliance obligations for the parties, but it is not automatically a generally binding regulation.

Formative check: Recast “a platform harms teenagers” into two narrower theories. One possible UDAP theory would allege that a specific representation or business practice was deceptive or unfair under a particular state statute; the evidence would need to address the representation, disclosure, practice, and statutory connection. One possible privacy theory would allege that a covered entity collected, used, disclosed, or retained specified information in a way prohibited by a particular privacy statute; the evidence would need to address the information, conduct, definitions, exemptions, and enforcement mechanism. Neither theory is established merely by the original statement.

This sequence also keeps procedural categories separate. An allegation is a party’s asserted claim. A civil investigative demand (CID) or subpoena seeks information; it is not itself proof. Discovery is evidence gathering during litigation. An adjudicated finding is a court’s determination after applying legal standards to the record. A settlement, consent decree, or assurance of voluntary compliance is a negotiated outcome. It may impose important compliance obligations, and implementation may involve a compliance monitor or structural remedy if the agreement provides for one, but the existence of a settlement alone does not establish wrongdoing or resolve every disputed legal question.

The distinction matters for multistate enforcement. A multistate coalition is a group of state AG offices coordinating investigation or litigation, not a single new sovereign. Each state must still connect its action to its own authority, jurisdiction, elements, and remedy. Jordan Ellis might assume that a lawsuit joined by many states automatically creates national law. It does not. Coordinated action can nevertheless produce practical spillover: a national platform may adopt one product or compliance change rather than operate under many different state systems. That choice can create a de facto national policy without giving the states formal power to command the entire country. Polycentric governance helps explain how multiple institutions can exert coordinated pressure while retaining distinct legal roles. #(Ostrom, 2010, Beyond Markets and States: Polycentric Governance of Complex Economic Systems)

Federal delay can create a regulatory gap when public concern is substantial but existing law does not cover the conduct, does not authorize the relevant actor, or does not provide a suitable remedy. Delay does not remove legal constraints; it may instead increase reliance on narrower state statutes, common-law theories, investigations, or negotiated outcomes. This is why the seriousness of a concern and the availability of a cause of action are related but different questions. The concern may motivate action, while the cause of action determines what can formally be alleged and proved.

Conclusion

Turning public concern into a legal claim requires more than showing that a practice appears harmful. The analyst must identify specific conduct, select a legal source, test jurisdiction-specific elements, gather evidence, establish standing or enforcement authority, and match the requested remedy to the legally recognized injury. UDAP laws focus on unfairness or deception; state privacy laws address defined data practices; common-law authority depends on recognized claims and local elements; federally authorized enforcement depends on the limits of federal statutes such as COPPA provisions or Clayton Act section 4C; and parens patriae protects certain statewide interests without becoming unlimited representation of private injuries. This framework explains both the possibilities and limits of state technology enforcement. It prepares the regulatory-gap analysis in CH1-SC3 and supplies the claim vocabulary needed for multistate coalition strategy in CH2-SC1: formal legal authority may be fragmented, yet coordinated enforcement can still produce significant national practical influence.

1.3: The Regulatory Gap Is Not a Legal Blank Check

Learning objectives

• Define a regulatory gap as a mismatch between rapid technological change and available federal action.
• Compare legislation, rulemaking, administrative enforcement, litigation, and settlement as different governance mechanisms.
• Explain why federal delay can motivate state action while leaving preemption, jurisdiction, and remedy questions unresolved.

Suppose the fastest way to change a platform’s code is a negotiated commitment, while the slowest route is a statute enacted through a legislature. Which route carries the stronger democratic mandate? The answer matters because a regulatory gap does not mean that law has disappeared. It means that technological or social change has moved faster than federal legislation or agency action. As the previous subchapter established, state action still requires a specific legal basis. The remaining institutional puzzle is timing: when federal action moves slowly, which mechanism can respond, and what is gained or lost when enforcement rather than legislation becomes the vehicle for change? Governance often operates through overlapping institutions rather than one complete rulebook #(Baldwin, Cave, and Lodge, 2012, Understanding Regulation: Theory, Strategy, and Practice).

Five Ways Institutions Can Shape Platform Conduct

The five mechanisms below are not interchangeable. They differ in who acts, what formal authority exists, how quickly conduct may change, and whether the result binds broadly or only in a particular dispute.

%FULL[infograph] %A comparison table rendered as a visual flow showing legislation, rulemaking, administrative enforcement, litigation, and settlement by primary actor, speed, scope, and type of obligation.

Mechanism	Main actor and formal authority	Typical speed and scope	Form of obligation	Democratic mandate and durability	Main limitation
Legislation	Federal or state legislature enacts a statute within its constitutional authority	Usually slower; potentially broad and generally applicable	Statutory requirement	Direct legislative mandate; may be durable, but can be amended, repealed, or challenged	Requires political agreement and remains subject to federal preemption and constitutional limits
Rulemaking	Federal or state agency acts under delegated statutory authority and required procedures	Medium to slow; generally applies to covered entities	Agency rule with legal force if valid	Indirect mandate through legislation and agency process; may last until changed or set aside	No valid rulemaking without sufficient delegation, procedure, and statutory authority
Administrative enforcement	Agency or authorized State attorney general investigates and applies existing law	Often faster; usually targeted to conduct or entities within authority	Warning, demand, order, or other enforcement measure, depending on context	Derives authority from existing law; usually narrower than a new statute	A warning is not automatically a binding enforcement action, and enforcement cannot create an entirely new general duty
Litigation	State attorney general, agency, private party, or other authorized plaintiff asks a court to decide claims	Variable and often slower; limited by claims, parties, jurisdiction, and judgment	Court judgment and authorized remedy	Judicially adjudicated result; may clarify law, but is subject to appeal and does not automatically bind every platform	Allegations in a complaint are not adjudicated findings
Settlement	Plaintiff and defendant negotiate; a court may approve the result	Often faster than trial; focused on named parties and agreed terms	Negotiated compliance obligation	Commitment depends on the agreement; a consent decree may be court-enforceable, while an assurance of voluntary compliance depends on applicable law and terms	A settlement is not legislation and ordinarily does not establish universal precedent

Legislation changes the law itself. A legislature can establish a generally applicable requirement, subject to constitutional authority and any federal preemption. Its procedures may be slow because proposals must move through political negotiation, voting, and executive processes. That slower pace can provide a stronger direct democratic mandate than a private agreement.

Rulemaking allows an agency to fill in regulatory details when a statute has delegated sufficient authority. An agency rule may apply broadly to a category of regulated conduct, but the agency must stay within its statutory authorization and follow applicable procedures. A rulemaking therefore has more general reach than a single enforcement action, but it does not have authority simply because an agency believes a policy would be useful.

Administrative enforcement applies existing legal requirements. A federal agency, state agency, or State attorney general may investigate, request information, warn, or pursue a violation when authorized. A civil investigative demand (CID) or subpoena may seek information; it is not itself a final finding that misconduct occurred. A warning may communicate an agency’s concern or invite compliance, but its legal effect depends on the authority and procedural setting.

Litigation uses a court to test legal claims and request a remedy. Filing a lawsuit begins a process; it does not create a national rule. A complaint contains allegations. Evidence may support those allegations, but only a judicial determination produces an adjudicated finding. The resulting remedy depends on the claims, parties, jurisdiction, and judgment.

Settlement is a negotiated resolution. It can require significant product or operational changes, but the obligation normally rests on the agreement and the parties. A consent decree is a settlement entered as a court order. An assurance of voluntary compliance is a negotiated undertaking whose enforceability depends on applicable law and its terms. Neither form is automatically a statute. Settlement can change conduct without deciding whether every allegation is legally correct, a concern explored in debates about negotiated justice #(Fiss, 1984, Against Settlement).

The eight-stage enforcement pathway used in this course helps keep these categories separate. It can be read as a movement from a concern, through authority and evidence, coordination and mechanism choice, to litigation or negotiation, remedy or commitment, and implementation or review. This is an analytical map, not a universal mandatory procedure; jurisdiction-specific requirements differ.

flowchart LR
  n1["Reported concern"] --> n2["Legal basis"]
  n2 --> n3["Evidence gathering"]
  n3 --> n4["Coordination"]
  n4 --> n5["Mechanism choice"]
  n5 --> n6["Litigation or settlement"]
  n6 --> n7["Remedy or commitment"]
  n7 --> n8["Implementation and review"]

The central comparison is therefore not simply “fast” versus “slow.” A fast settlement may change a defendant’s operations quickly, while a statute may create a broader and more durable rule. Litigation may produce an adjudicated finding, while enforcement or settlement may resolve conduct without a final judicial ruling. Instrument choice determines both what changes and what kind of authority the change represents.

Why Federal Delay Invites State Action

A regulatory gap appears when federal legislation or agency action does not keep pace with a changing technology or social problem. Federal gridlock helps explain why this can happen, but gridlock has several causes: political polarization, divided leadership, limited statutory authority, budget constraints, slow legislative processes, and judicial pushback. Legislative stalemate is often the result of accumulated institutional veto points rather than one obstacle #(Binder, 2003, Stalemate: Causes and Consequences of Legislative Gridlock) #(Jones, Sulkin, and Larsen, 2003, Policy Punctuations in American Political Institutions).

Federal delay can motivate states because states may still possess concurrent authority. Federalism is not a simple rule that one level of government acts only when the other is silent. It is an arrangement in which federal and state institutions may address related problems at the same time, subject to constitutional and statutory limits #(Bednar, 2008, The Robust Federation: Principles of Design). State police powers, UDAP laws governing unfair or deceptive acts or practices, state privacy laws, common-law authority, and parens patriae authority may provide possible legal bases. The precise authority of a State attorney general or state agency remains jurisdiction-specific.

The institutional roles are different:

• A state legislature can enact a state statute, creating a new legal requirement within its authority.
• A State attorney general may investigate and enforce existing laws, use a CID or subpoena where authorized, bring litigation, or negotiate a settlement. An AG cannot create a statute merely by announcing a policy.
• A state agency may issue rules when a legislature has delegated authority and may conduct administrative enforcement.
• A court can adjudicate claims and grant remedies authorized by law. It does not automatically convert a lawsuit into a nationwide legislative command.
• A multistate coalition can coordinate investigations or litigation among several states. Coordination can increase resources and leverage, but it does not merge state authority into federal authority.

The sequence can be summarized as follows:

flowchart TD
  n1["Rapid technological change"] --> n2["Federal delay"]
  n2 --> n3["Regulatory gap"]
  n3 --> n4["State investigation or action"]
  n4 --> n5["Platform operational response"]
  n5 --> n6["Practical spillover"]
  n2 --> n7["Existing legal constraints remain"]
  n7 --> n8["Preemption, jurisdiction, remedy, and constitutional review"]

The most important distinction is between formal legal authority and practical national influence. Formal legal authority asks: who is legally bound, by what instrument, under which law, and within what jurisdiction? Practical national influence asks: did the platform change its nationwide operations because maintaining several state-specific systems was costly or impractical?

For example, imagine that a platform reaches a settlement with a group of states and then applies the agreed operational change across the country. The change may function as a de facto national policy through the platform’s implementation choice. It may create practical spillover beyond the states that acted. Yet the settlement’s formal obligation may still apply only to the defendant and according to the agreement’s terms. It is not binding national legislation or universal judicial precedent. Whether a platform actually adopts a nationwide system is an empirical question, not a legal presumption. Regulatory diffusion research describes how one jurisdiction’s rules can influence conduct elsewhere without formally governing every actor #(Vogel, 1995, Trading Up: Consumer and Environmental Regulation in a Global Economy) #(Bradford, 2012, The Brussels Effect).

Federal delay also does not erase limits. Federal preemption may displace state law when federal law expressly or implicitly overrides it, or when state requirements conflict with federal authority. Jurisdiction concerns whether a court or agency has power over the parties, conduct, or dispute. The Dormant Commerce Clause may limit state measures that discriminate against or excessively burden interstate commerce, but whether it does so depends on the law’s design and effects. The First Amendment and Section 230 may also become relevant depending on the claim; their presence does not resolve every dispute in advance.

Finally, a remedy must fit the legal basis and proceeding. A requested operational change, structural remedy, or compliance monitor requires appropriate authority or agreement. The fact that a state remedy affects a national product does not by itself make the remedy unconstitutional. National effects may matter to the analysis, but they do not replace examination of preemption, jurisdiction, constitutional principles, and the remedy’s actual design.

Conclusion

A regulatory gap is a timing problem, not a legal blank check. Federal gridlock may slow legislation or agency action while leaving federal authority, state authority, constitutional limits, and judicial remedies in place. The five governance mechanisms provide different responses: legislation makes general law; rulemaking develops delegated regulatory requirements; administrative enforcement applies existing law; litigation seeks judicial resolution; and settlement creates negotiated commitments. Their speed, scope, democratic mandate, and durability differ.

A compact formative classification check captures the distinctions: a new statute is legislation, an agency rule is rulemaking, a warning may be administrative communication but is not automatically completed enforcement, a filed lawsuit is litigation, and negotiated product reform is settlement. The privacy-law question has the same answer: if Congress has not enacted a comprehensive national privacy law, states cannot therefore impose any privacy requirement they choose. The relevant analysis still asks which legal basis applies, whether federal preemption or constitutional limits intervene, whether the forum has jurisdiction, and whether the requested remedy is authorized. These distinctions prepare the ground for Chapter 2, where coalition-building becomes a strategy for converting dispersed state authority into coordinated pressure, and for the final chapter, which evaluates outcomes by separating formal legal authority from practical national influence.

Chapter 2: From Separate Offices to Collective Leverage

2.1: Collective Action Against Resource Asymmetry

Chapter 1 established that a state must identify a specific legal basis before acting. That authority does not eliminate the practical disadvantage of confronting a national platform alone. Jordan’s central question is therefore precise: if a state joins other states, does it gain new legal authority, or does it mainly improve its bargaining position? The answer is usually the latter. Through federalism, states retain separate sovereign roles while coordinating enough capacity to investigate, litigate, negotiate, and monitor conduct that crosses state borders.

Learning objectives: By the end of this subchapter, you should be able to:

• Explain why a single state may face resource, jurisdictional, and fragmented-impact disadvantages.
• Describe how a multistate coalition organizes legal, technical, investigative, and political capacity.
• Evaluate how bipartisan leadership can broaden political support without proving that allegations are true.

Why One State May Struggle Alone

In a federal system, authority is distributed among state and federal governments. A State attorney general may rely on state police powers, which are powers to protect public health, safety, welfare, and consumers. Depending on the jurisdiction, additional legal bases may include UDAP laws governing unfair or deceptive acts or practices, state privacy laws, common-law authority, or parens patriae authority to protect certain collective state interests. Whether a particular claim is available remains jurisdiction-specific.

A single-state leverage gap is the difference between possessing formal legal authority and possessing enough practical capacity to obtain information, sustain litigation, and secure a meaningful remedy against a much larger opponent. The gap does not mean that a single-state action is ineffective. It means that the state may confront disadvantages that affect timing, cost, evidence, and bargaining power.

Three disadvantages commonly interact:

• Resource asymmetry: A large technology platform may deploy extensive legal teams, technical experts, economists, data analysts, and financial resources. A state legal division may have fewer personnel and less capacity to sustain a lengthy, technically complex case. Resource asymmetry can affect discovery management, expert access, and the speed of investigation. It does not determine the legal outcome; a State attorney general may possess civil investigative demands, subpoenas, consumer complaints, and public-interest evidence unavailable to private litigants. Research on repeat players illustrates how experience and resources can create structural litigation advantages, while also recognizing that public actors possess distinctive powers #(Galanter, 1974, Why the “Haves” Come Out Ahead: Speculations on the Limits of Legal Change).

• Jurisdictional friction: A platform may operate through systems, employees, contracts, and data located across several jurisdictions. A state must connect the challenged conduct and resulting injury to its legal authority. The platform may also raise federal preemption arguments, meaning that federal law displaces state law, or constitutional arguments under the Dormant Commerce Clause or the First Amendment. Section 230 may also be relevant to some platform-related legal arguments. The applicability and strength of these arguments depend on the claim and the court.

• Fragmented geographic impact: Harm may be experienced by users in many states, but no single state necessarily possesses the complete factual picture. Complaints, affected groups, and relevant records may be dispersed. A single office may have difficulty distinguishing a local pattern from a broader one.

These disadvantages concern practical leverage rather than automatic legal weakness. A state with strong evidence, a clear statute, and a carefully tailored remedy may succeed alone. Conversely, a large coalition may still have weak claims if the evidence is inadequate or the law does not support the requested remedy. A nationwide product change also does not necessarily exceed state remedial authority. A platform may voluntarily adopt one system nationwide because maintaining different systems is costly. That practical spillover should not be confused with a state issuing a formally binding national rule.

Designing a Coalition Without Creating a New National Regulator

A multistate coalition is an arrangement in which several states coordinate investigation, experts, pleadings, litigation strategy, or settlement strategy while retaining separate sovereign claims. Each state remains responsible for its own legal authority, procedural requirements, and final decisions. The coalition increases capacity without becoming a federal agency.

This arrangement is usually a form of informal intergovernmental collaboration, not a congressionally approved interstate compact. An interstate compact is a formal agreement among states that may require specific legal approval. Informal collaboration is more flexible, but it can produce differences in confidentiality rules, information-sharing practices, staffing, governance, and settlement approval. Coalition members may also have different statutes, injury patterns, litigation objectives, or preferred remedies. Cooperation does not erase those differences.

%FULL[infograph] %A network diagram showing an executive committee, legal strategy team, technical and data investigators, and general coalition members, with arrows indicating information and resource sharing.

A coalition may distribute work through several roles:

• An executive committee or co-leads may set priorities and communicate with other offices.
• A legal strategy team may compare statutes, develop pleadings, coordinate motions, and assess litigation risk.
• Technical and data investigators may analyze product design, internal processes, records, and consumer experiences.
• General coalition members may contribute complaints, local evidence, staff time, legal theories, or views about remedies.

The resulting structure reflects polycentric governance: multiple centers of authority coordinate around a shared problem without merging into one institution #(Ostrom, 1990, Governing the Commons). It also addresses collective-action problems. Sharing experts and investigative costs can reduce duplication, although states may disagree about priorities or benefit from work performed by others without contributing equally #(Olson, 1965, The Logic of Collective Action).

The eight-stage enforcement pathway shows how dispersed authority can become coordinated practical pressure. The stages are analytical rather than rigid. Investigation and negotiation may overlap, and litigation may begin before settlement discussions conclude.

Stage	Authority and objective	Main tool	Timing	Principal limitation
1. Identify harm	Determine whether reported conduct may affect consumers or other public interests	Complaints, public records, preliminary inquiries	Before formal action	Reports and allegations are not yet proof
2. Assess authority and gaps	Match the suspected conduct to available state law and identify regulatory gaps	Legal research and jurisdictional analysis	Early assessment	Statutory limits, preemption, and constitutional defenses may apply
3. Pool resources	Reduce the single-state leverage gap	Shared staff, experts, funding, and expertise	After common concerns emerge	Participation and contributions may differ
4. Coordinate investigation	Build a fuller factual record across jurisdictions	Civil investigative demands, subpoenas, interviews, and data protocols	Before or alongside litigation	Confidentiality and information-sharing rules vary
5. Litigate	Present legally sufficient claims to a court	Complaints, motions, discovery, and expert testimony	When formal adjudication or litigation pressure is needed	Courts control procedure, evidence, and legal interpretation
6. Negotiate remedies	Resolve uncertainty and seek corrective obligations	Settlement, assurance of voluntary compliance, or consent decree	During or after investigation or litigation	A negotiated resolution is not automatically an adjudicated finding
7. Secure enforceability	Make commitments specific enough to implement and enforce	Court-approved orders, reporting duties, deadlines, or other specified terms	At resolution	Vague or weakly authorized commitments are harder to enforce
8. Monitor compliance	Determine whether obligations are being carried out	Reports, audits, follow-up inquiries, or a compliance monitor where provided	After resolution	Monitoring requires continuing resources and cooperation

The pathway can be summarized as a movement from capacity to coordination and then to implementation:

flowchart LR
  n1["Single-state leverage gap"] -->|"pool capacity"| n2["Multistate coalition"]
  n2 -->|"coordinate evidence and pleadings"| n3["Separate sovereign actions"]
  n2 -->|"use co-leads from both parties"| n4["Broader political signal"]
  n2 -->|"manage through informal collaboration"| n5["Negotiation and monitoring"]
  n5 -->|"may create practical spillover"| n6["Broader product or compliance effect"]

Suppose, in a bounded hypothetical, several states receive similar complaints about a platform practice. Jordan should not conclude that the coalition has created national law. The states still must establish their own claims. The coalition may, however, make it easier to compare evidence, retain technical experts, coordinate pleadings, and present the platform with greater settlement exposure. This is a change in bargaining position, not necessarily a change in formal legal authority. State-led coordination can nevertheless contribute to national policymaking in practice #(Nolette, 2015, Federalism on Trial: State Attorneys General and National Policymaking in Contemporary America).

The categories used to describe the case must remain separate:

• An allegation is a claim asserted by an investigating office or party.
• Evidence is information offered to support or contest that claim; its legal significance may depend on admissibility, reliability, and context.
• A settlement records negotiated commitments. It may resolve uncertainty without deciding whether the alleged violation occurred.
• A regulatory action is an action taken under delegated regulatory authority, such as rulemaking or administrative enforcement. Coalition litigation and settlement are not automatically regulation.
• An adjudicated finding is a court’s determination after applying legal standards to the record. A settlement, coalition size, or bipartisan participation does not substitute for such a finding.

Both sides may sometimes prefer settlement, but neither side must do so. States may seek an enforceable remedy while reducing the cost, delay, and risk of trial. A platform may prefer predictable obligations to prolonged litigation or inconsistent state-by-state systems. Enforceability depends on the legal instrument, the specificity of the commitments, the scope of the remedy, reporting requirements, monitoring arrangements, and available consequences for noncompliance. A structural remedy, which changes an organization’s systems or practices rather than merely providing compensation, may require especially clear implementation and monitoring terms.

Bipartisan leadership illustrates the difference between political breadth and factual proof. Coalitions may use co-leads from both major political parties to signal that the action is not exclusively associated with one party. That signal may broaden public or institutional support, but it does not prove liability, legal sufficiency, or the truth of the allegations. Elected State attorneys general may still respond to constituent preferences, public visibility, and electoral incentives. Legitimacy and legal merits are separate categories, a distinction developed in institutional legitimacy research #(Suchman, 1995, Managing Legitimacy: Strategic and Institutional Approaches). Chapter 3, Subchapter 3 returns to these political incentives and their institutional limits.

Conclusion

Collective action addresses resource asymmetry by pooling capacity without pooling sovereignty. The single-state leverage gap arises when limited resources, jurisdictional friction, and geographically dispersed effects make unilateral action more difficult, not when a state necessarily lacks authority. A multistate coalition can coordinate experts, evidence, pleadings, litigation, settlement strategy, and monitoring while each State attorney general preserves separate legal claims and decision-making power.

Coalition size is therefore only one possible source of leverage. A careful assessment should weigh resource pooling, shared evidence, nationwide settlement exposure, and political breadth according to the case’s principal bottleneck. A politically broad coalition with little technical expertise may gain visibility and coordination but still struggle to explain complex platform conduct or design a workable remedy. Conversely, strong technical evidence without effective governance may produce internal disagreement. Informal collaboration creates flexibility, but confidentiality, information-sharing, free-riding, and differing remedies require active management #(Feiock, 2013, The Institutional Collective Action Framework). The next subchapter develops the investigation pathway in greater detail, showing how legal, technical, and operational capacity is assembled before claims are tested in litigation or translated into enforceable compliance obligations.

2.2: Evidence First—Investigation, Filing, and Litigation Risk

A case file grows from an allegation into a contest over documents, technical design, jurisdiction, and legal characterization. Coalition design answers who works together; it does not yet answer how a suspicion becomes a defensible case. The central question is therefore evidence-first: what must states learn, test, and preserve before choosing litigation or settlement?

The enforcement pathway has eight stages: perceived harm, legal-authority matching, leverage assessment, coalition formation, investigation, litigation or settlement choice, compliance, and practical spillover. The first six are the focus here. An allegation is an assertion of possible misconduct; evidence is material that supports or challenges that assertion; an adjudicated finding is a court’s determination; and a negotiated commitment is an obligation accepted through settlement. Moving from one category to another requires more than collecting documents.

%FULL[infograph] %An eight-stage circular process from perceived harm to legal authority, leverage gap, coalition, investigation, litigation or settlement choice, compliance, and spillover, with the first six stages emphasized.

From Perceived Harm to Coordinated Discovery

The first six stages can be organized as a sequence of changing institutional questions:

Stage	Central question	Evidence or action	Principal limitation
1. Perceived harm	What appears to be happening?	Complaints, reports, or preliminary indicators	A concern is not yet a legal violation
2. Legal-authority matching	Which institution has authority?	Possible UDAP, privacy, common-law, or other theories	Authority varies by state and claim
3. Leverage assessment	What can one state realistically accomplish?	Assessment of information, resources, and platform scale	Practical influence is not formal power
4. Coalition formation	Who will coordinate?	Shared roles, experts, and investigative resources	States remain legally separate
5. Investigation	What facts can be established or tested?	CIDs, subpoenas, discovery, and technical analysis	Obtained information may be disputed or incomplete
6. Filing or settlement choice	Which forum and remedy best fit the record?	Complaint, negotiated resolution, or both	Neither filing nor settlement guarantees a legal result

Stage 1: perceived harm. A state attorney general (AG) may begin with consumer complaints, public reporting, or other indications of harm. At this point, the material identifies a problem to investigate. It does not establish that the platform caused the harm, violated a statute, or can be compelled to adopt a particular remedy.

Stage 2: legal-authority matching. The AG’s office next identifies a source of formal legal authority: the power law grants an institution to investigate, sue, or seek relief. Depending on state law, possible sources may include state police powers, UDAP laws prohibiting unfair or deceptive acts, state privacy laws, common-law authority, and parens patriae authority to protect specified public interests. Federalism—the allocation of governing authority between state and federal institutions—means that these powers may overlap with federal authority without becoming identical to it.

Stage 3: leverage assessment. A single state may face a leverage gap against a national platform. The platform may possess greater technical information, financial resources, and ability to spread a challenged practice across jurisdictions. A state may have formal legal authority but limited practical capacity to obtain and interpret the necessary evidence. This gap helps explain why an AG may seek a multistate coalition.

Stage 4: coalition formation. A coalition pools personnel, technical expertise, and litigation resources. It can coordinate questions, divide investigative work, and reduce duplicative demands. It does not become a single national regulator. Each AG retains separate legal authority, institutional interests, and responsibility for decisions. This distinction reflects the idea of “uncooperative federalism”: states can influence national policy through their own legal institutions while remaining institutionally distinct from the federal government #(Bulman-Pozen & Gerken, 2009, “Uncooperative Federalism”).

Stage 5: investigation. A civil investigative demand (CID) is a state investigative tool that, where authorized, may seek internal documents, design specifications, user metrics, communications, and testing records. The exact authority, scope, and enforcement procedure differ among states. A CID therefore reduces an information gap; it does not prove the facts requested by the state. Treating investigative material as conclusive proof of liability confuses investigation with adjudication.

A subpoena is a compulsory request for documents, testimony, or other information issued under applicable legal authority. Its precise use and limits depend on the governing law and procedural setting. Discovery is the formal evidence-gathering process associated with litigation. It may allow parties to obtain information needed to evaluate product features, internal knowledge, communications, testing, and possible links between conduct and injury. The important point is functional: these tools help build a record for evaluation. They do not automatically resolve relevance, privilege, authenticity, admissibility, or interpretation.

When several jurisdictions work from the same materials, coordinated discovery becomes an operational system rather than merely a legal event. A joint document repository allows AG offices and expert teams to organize shared materials, reduce repeated requests, and compare the platform’s representations across jurisdictions. Experts may help explain technical systems, metrics, or testing records. Administrative investigation is therefore not only a prelude to decision; it is also an information-gathering institution #(Mashaw, 2012, “The Administrative Constitution”).

Coordination creates its own governance obligations. The coalition must address:

• Privilege: A document may contain protected attorney-client communications or attorney work product. Sharing it incorrectly may create disputes about protection or waiver.
• Confidentiality: Internal business information, personal information, and technical material may require access controls.
• Data security: A repository holding sensitive platform or user information must be managed so that coordination does not create a new security risk.
• Public-records differences: Materials treated as confidential in one jurisdiction may be subject to different disclosure rules elsewhere.
• Interpretation and admissibility: A shared document is not automatically admissible, and several offices may reasonably disagree about what it proves.

Collective-action theory helps explain the tradeoff. Shared systems can reduce duplication, but cooperation requires rules for information sharing, monitoring, and conflict management #(Ostrom, 1990, Governing the Commons). A joint repository creates practical coordination, not centralized legal authority.

Stage 6: filing or settlement choice. After investigation, the coalition can evaluate whether the record supports litigation, negotiation, or a combination of approaches. The choice should reflect not only the volume of documents but also the strength of the factual interpretation, the likely defenses, the available forum, and the feasibility of a remedy. Thus, evidence changes the case at every stage: it moves the matter from concern, to authority, to coordinated inquiry, to a defensible strategic choice.

Choosing a Filing Path and Testing the Defenses

States may pursue a centralized federal complaint, parallel state actions, or both. A court’s jurisdiction is its authority to hear a dispute and issue binding orders. Federalism and jurisdiction make forum selection consequential: a federal filing and state filings do not produce identical procedural consequences. Federal preemption—the displacement of state law by federal law—and potential Dormant Commerce Clause limits on state action affecting interstate commerce may also shape the dispute. Their application is claim- and jurisdiction-specific.

%HALF left[infograph] %A side-by-side pathway diagram comparing one centralized federal complaint with parallel state actions, showing differences in forum, coordination, procedural control, and inconsistency risk.

Pathway	Forum and authority	Objective	Procedural control	Main limitation
Centralized federal complaint	One federal proceeding, if jurisdiction permits	Coordinate claims and evidence	More centralized case management	Consolidation may create jurisdictional and strategic tradeoffs
Parallel state actions	Separate state proceedings under participating states’ authority	Preserve state-specific theories and forums	Each state retains greater control over its case	Duplication and inconsistent results remain possible
Combined approach	Federal and state proceedings pursued together or in sequence	Preserve options while increasing pressure	Coordination must account for multiple forums	Greater cost, overlap, and conflict risk

A centralized federal approach may make coordinated discovery and expert presentation more efficient. It may also place important questions within one procedural structure whose jurisdictional requirements, available claims, and rulings affect the participating states. Consolidation is not automatically beneficial.

Parallel state cases can preserve state-specific legal theories and allow each AG to use the forum associated with that authority. They can also produce different interpretations, schedules, or outcomes. A coalition may coordinate factual work without guaranteeing consistent judgments. The combined approach can preserve flexibility, but it multiplies procedural demands and increases the need for careful coordination.

The platform’s defenses often turn on legal characterization: what, exactly, is the state challenging?

Section 230 is a federal statute that may protect claims treating a platform as the publisher or speaker of third-party content. It does not necessarily immunize claims about the company’s own conduct, product design, commercial representations, or data practices. But a label does not decide the issue. Calling a claim “product design” does not automatically defeat a Section 230 defense if the substance of the claim still depends on treating the platform as responsible for publishing or arranging third-party material. The court must examine the actual allegations, challenged function, and requested relief. Platform scholarship similarly emphasizes the relationship among algorithmic functions, editorial activity, and claims directed at the platform’s own behavior #(Balkin, 2018, “Free Speech in the Algorithmic Society”).

The First Amendment may supply a related but distinct defense. A platform may argue that ranking, recommendation, and moderation involve protected editorial judgment. That argument does not mean every algorithmic operation is protected speech. A technical system may also process data, set defaults, design interfaces, make commercial representations, or implement product features. The First Amendment defense therefore does not necessarily resolve claims directed at product design, data practices, or representations. Platform-governance research underscores that moderation combines expressive judgments with technical and organizational practices #(Gillespie, 2018, Custodians of the Internet).

A disciplined analysis separates these questions rather than deciding them by labels:

flowchart TD
  n1["Challenged platform conduct"] --> n2["Publisher or speaker question"]
  n1 --> n3["Editorial judgment question"]
  n2 --> n4["Section 230 analysis"]
  n3 --> n5["First Amendment analysis"]
  n4 --> n6["Claim-specific proof"]
  n5 --> n6
  n6 --> n7["Causation, injury, and remedy"]

The next test is causation and injury. Causation asks whether the challenged conduct legally caused the relevant harm. Injury asks whether the claimant or protected interest suffered a harm recognized by the applicable claim. Both are claim-specific. Evidence that social-media use correlates with adverse mental-health outcomes may support further inquiry, but correlation alone does not establish that a particular product feature caused a particular legally recognized injury. Causal reasoning separates association from legally relevant attribution #(Hart & Honoré, 1985, Causation in the Law).

The evidence must therefore support a connected chain: challenged conduct, exposure or mechanism, injury, legal recognition, and a remedy that addresses the injury. Multiple contributing causes, uncertain measurements, competing expert interpretations, and differences among claimants can weaken that chain. A strong discovery record may reveal what the platform knew or tested, yet still leave disputes about whether the conduct caused the legally relevant injury and whether the requested remedy fits it.

These uncertainties shape litigation risk and settlement calculus. A settlement is a negotiated resolution rather than an adjudicated finding. States may prefer it when litigation is costly, defenses create uncertainty, and a concrete compliance obligation is more valuable than an uncertain judgment. A platform may prefer it to reduce expense, litigation exposure, and operational uncertainty. A settlement does not, without more, establish liability or create binding precedent. Its practical value depends on the legal vehicle and the precision of its terms.

An enforceable remedy generally requires clearly stated obligations, measurable standards, reporting or verification, a defined duration, and a mechanism for responding to noncompliance. A consent decree is a court-approved resolution; an assurance of voluntary compliance is a negotiated commitment whose force depends on applicable law and its terms. A compliance monitor may oversee implementation when the resolution provides for one. A structural remedy may address systems, processes, or organizational practices rather than merely compensate for past harm. These choices are fact- and jurisdiction-dependent.

Conclusion

The evidence-first pathway converts dispersed authority into coordinated practical pressure through a sequence, not a shortcut. Perceived harm leads to legal-authority matching; the leverage gap encourages coalition formation; CIDs, subpoenas, discovery, shared repositories, and experts develop a record; and that record informs the choice among federal, state, parallel, or negotiated paths. Coordination can reduce duplication without creating a national regulator, while Section 230 and First Amendment arguments require claim-specific analysis rather than categorical assumptions. Causation and legally recognized injury remain independent proof requirements.

The final two stages are compliance and spillover. A validly imposed or negotiated compliance obligation can change platform practice, and a national platform may extend that change beyond the legally required jurisdictions, producing practical spillover or a de facto national policy. That practical influence is not the same as formal legal authority. These distinctions provide the factual and procedural bridge to the settlement calculus in the next subchapter and the legal-defense vocabulary needed to read the later Meta case without confusing allegations, evidence, negotiated commitments, and adjudicated findings.

2.3: The Bargain After the Threat—Remedies, Monitoring, and Spillover

Investigation and filing increase pressure, but pressure alone does not protect consumers. The process now turns to a strategic choice: should the parties continue toward trial, or negotiate a remedy that changes the platform’s conduct sooner? Maya Rivera frames the central problem precisely: Is a settlement an admission, a compromise, or a regulatory instrument? The answer depends on the settlement’s text, its enforcement mechanism, and its practical reach. By the end of this subchapter, readers should be able to evaluate that choice, distinguish major remedy forms, and explain how a state-level agreement can influence users nationally without becoming national legislation.

A single State attorney general (AG) may face a leverage gap when challenging a national platform. The platform may have greater financial resources, centralized technical systems, and the ability to prolong litigation. Federalism—the division of governmental authority among national and state institutions—allows states to exercise concurrent authority over some conduct, but it does not make one state a national regulator. A multistate coalition can pool legal authority, investigative capacity, and resources while remaining a coordinated group of state actors rather than a single national agency.

%FULL[infograph] %A balanced comparison visual showing trial and settlement branches, followed by remedy types, monitoring, and possible nationwide operational spillover without labeling the result as national legislation.

Why Both Sides May Settle

The eight-stage pathway in this subchapter is an analytical map rather than a universal procedural checklist. Particular matters may combine stages, repeat them, or proceed differently. The pathway nevertheless clarifies how pressure can become an operational remedy:

1. Identify suspected harm. A state or group of states identifies conduct that may violate applicable law. At this stage, the matter involves an allegation, not an adjudicated finding.
2. Pool resources. State AGs form or join a multistate coalition to coordinate personnel, legal theories, discovery, and strategy.
3. Investigate. States use tools such as civil investigative demands (CIDs), subpoenas, interviews, and coordinated discovery to develop evidence.
4. File and litigate. Pleadings formalize claims and create litigation pressure, while the parties contest legal theories and factual interpretations.
5. Bargain or continue toward trial. The parties compare the value of a negotiated resolution with the risks of adjudication.
6. Specify remedy terms and legal instrument. If they settle, the agreement identifies payments, conduct restrictions, product changes, disclosures, reporting duties, or other obligations.
7. Implement the remedy. The platform translates legal language into product, software, data, advertising, and organizational changes.
8. Monitor compliance. States, courts, or an appointed monitor assess whether the agreed reforms have been carried out.

flowchart LR
  n1["Identify suspected harm"]
  n2["Pool state resources"]
  n3["Investigate"]
  n4["File and litigate"]
  n5["Bargain or continue to trial"]
  n6["Specify remedy terms"]
  n7["Implement reform"]
  n8["Monitor compliance"]
  n1 --> n2
  n2 --> n3
  n3 --> n4
  n4 -->|"creates pressure"| n5
  n5 -->|"settlement path"| n6
  n6 --> n7
  n7 --> n8

The important evidentiary distinction runs through all eight stages. Identification produces an allegation. Investigation may produce evidence. A trial can produce an adjudicated finding if a court resolves the dispute. A settlement, by contrast, produces negotiated commitments. A platform’s agreement to change a product does not, by itself, establish that the original allegation was true.

For a platform, trial may offer the possibility of defeating the claims, preserving legal arguments, or obtaining a favorable precedent. But trial also creates exposure to an adverse judgment, brand damage, financial uncertainty, appeals, litigation expense, and a precedent that could affect future disputes. Settlement may reduce those uncertainties and give the platform more control over the wording, timing, and design of the obligations. It can also produce faster consumer-facing changes than a lengthy trial. These advantages do not mean that a platform will always settle. It may continue litigating to avoid costly reforms, protect a contested legal position, or test the limits of state authority.

States face a different but equally difficult calculation. Settlement can deliver concrete reforms and monetary relief without waiting through trial and appeal. Yet states may give up an opportunity to obtain a judicial ruling, establish precedent, clarify disputed authority, or secure a remedy that reaches conduct more broadly. A settlement that is too vague or limited may fail the public interest even if it ends the case efficiently. Settlement theory therefore asks whether negotiated resolution serves affected people or merely removes a dispute from public adjudication #(Fiss, 1984, Against Settlement).

Path	Platform-side objective	State-side objective	Timing and uncertainty	Principal limitation
Trial	Contest liability, preserve legal arguments, and seek a favorable ruling	Obtain an adjudicated finding, precedent, or broader judicial remedy	Often longer, more expensive, and subject to appeal	No assurance that the result will produce the desired operational reform
Settlement	Limit exposure and obtain more predictable obligations	Secure specified consumer relief and operational change	Greater near-term certainty, but dependent on negotiated terms	May compromise legal claims and usually does not create binding precedent
Continued negotiation after filing	Improve the terms before a final judicial determination	Increase specificity, scope, or enforceability of relief	Pressure may increase as evidence and costs accumulate	Agreement may fail if the parties value legal principle or cost avoidance more than compromise

The litigation-versus-settlement calculus is therefore institutional, not merely financial. A useful analysis identifies the likely benefits of trial, the costs and uncertainty of trial, the specificity and enforceability of the proposed settlement, and the public-interest value of obtaining reform sooner. Neither “settlement is efficient” nor “trial is principled” is sufficient on its own.

Designing and Monitoring an Operational Remedy

A settlement is a negotiated resolution in which the parties agree to specified terms instead of obtaining a full judicial determination after trial. Its legal effect depends on the agreement’s text and the mechanism used to enforce it. A settlement may contain civil penalties, restitution, disclosures, conduct restrictions, product changes, reporting duties, or several of these at once.

The instrument matters because not every settlement document has the same level of judicial involvement or enforceability.

Tool	Primary objective	Supervision or enforcement character	Important limitation
Settlement	Resolve the dispute through negotiated commitments	Depends on the agreement and governing law	Does not necessarily admit wrongdoing or establish precedent
Consent decree	Convert negotiated terms into a court-enforceable settlement order	Often requires judicial approval and specifies obligations enforceable through the approving court	Usually binds the parties covered by the order, not unrelated companies
Assurance of voluntary compliance	Resolve a state-law enforcement matter through an assurance	May operate without the same form of judicial supervision as a federal consent decree; enforceability depends on state law and the instrument	“Voluntary” does not necessarily mean unenforceable, and the term is not interchangeable with consent decree
Civil penalty	Sanction conduct and promote deterrence	Payment is directed to the government or another authorized public recipient	A payment may not change the underlying product or data practice
Restitution	Compensate consumers or return value to affected parties	Distribution depends on eligibility, proof, and governing law	Harm may be difficult to identify, quantify, or distribute
Structural remedy	Change operations, products, systems, or governance	Requires implementation standards and often continuing oversight	Feasibility, scope, cost, and unintended effects must be evaluated
Compliance monitor	Verify whether agreed reforms have been implemented	An independent technical auditor or outside monitor reports according to a defined mandate	Monitoring itself does not guarantee compliance

A consent decree is thus more than a settlement in the ordinary sense: it is a court-enforceable settlement order, often requiring judicial approval and specifying continuing obligations. A negotiated agreement does not become a consent decree merely because it resolves litigation. Nor does a consent decree become legislation. It ordinarily governs the parties subject to the order rather than unrelated companies.

An assurance of voluntary compliance is a state-law settlement instrument that may involve less judicial supervision than a federal consent decree. Its practical force depends on the relevant state law, the terms of the assurance, and available enforcement mechanisms. Because state law varies, readers should examine the actual instrument rather than assume that every assurance has the same legal consequences.

Structural remedies address future operations rather than only past conduct. Possible examples include changing an interface, revising privacy defaults, adopting age-assurance practices, limiting advertising, strengthening disclosures, or altering data-retention practices. These reforms can be significant because they require a platform to redesign how users encounter a service, how information is collected, how advertisements are delivered, or how internal systems retain data. But a product change does not prove that the original allegations were true. It may reflect compromise, risk management, a business decision, or an effort to avoid prolonged litigation.

Evaluation must therefore ask whether the remedy is feasible, sufficiently specific, appropriately scoped, measurable, and connected to the alleged harm. A reform may appear substantial while producing little practical change, or it may solve one problem while creating unintended effects elsewhere.

Monitoring connects legal language to operational evidence. Suppose a settlement requires a youth-safety product reform. A monitor would need, at minimum, information about:

• The precise reform required.
• The platform’s baseline condition before implementation.
• The technical or operational standard for determining compliance.
• Relevant product documentation, testing records, and system evidence.
• A method for detecting partial, inconsistent, or geographically limited implementation.
• Reporting deadlines, monitoring duration, and procedures for addressing deficiencies.

These are analytical criteria, not universal legal requirements. The actual mandate depends on the decree or assurance. A monitor’s effectiveness also depends on access to systems and records, independence, reporting duties, authority to identify deficiencies, and sufficient duration. A monitor may verify that a specified feature was implemented without being able to prove that the feature eliminated every alleged harm. Continuing oversight can also create governance problems if responsibilities are unclear or supervision becomes too managerial #(Resnik, 1982, Managerial Judges).

Monetary remedies complement but do not replace structural remedies. Civil penalties generally emphasize sanction and deterrence; restitution generally emphasizes compensation. Deterrence depends not only on the size of a payment, but also on the probability of detection, the certainty of enforcement, and the target’s ability to absorb or avoid the cost #(Polinsky & Shavell, 2000, The Economic Theory of Public Enforcement of Law). A financial penalty therefore cannot guarantee future compliance. Without implementation standards, monitoring, and consequences for noncompliance, payment may leave the underlying architecture unchanged.

Finally, operational change can extend beyond the states that negotiated the agreement. A platform may adopt one product or code change nationwide because maintaining many state-specific systems is costly or technically difficult. This is practical spillover: a consequence of implementation and business strategy, not a new source of formal legal authority. The platform may instead use geofencing or other technical tools to differentiate users by location. Every platform should not be assumed to lack that ability.

If a platform applies the same reform nationwide, the result may be described as a de facto national policy. It is not automatically national law, national legislation, or a binding rule for unrelated companies. The distinction turns on legal source, scope of obligation, enforceability, and practical reach. Multistate enforcement illustrates polycentric governance: multiple public actors coordinate overlapping authority without becoming one centralized regulator #(Ostrom, 2010, Beyond Markets and States: Polycentric Governance of Complex Economic Systems).

Conclusion

The bargain after litigation pressure is not a single event but a chain of choices. States identify suspected harm, pool resources, investigate, litigate, bargain, define a legal instrument, require implementation, and monitor compliance. Trial may produce an adjudicated finding and precedent, while settlement may produce faster and more specific reforms but ordinarily remains a negotiated compromise. A consent decree is a court-enforceable settlement order; an assurance of voluntary compliance is a distinct state-law instrument whose force depends on its terms and governing law. Civil penalties and restitution address sanction and compensation, while structural remedies seek to change operations. Compliance monitors test whether those changes occurred, but monitoring does not guarantee success. The coalition can formally command only what its authority and instrument reach—usually obligations imposed on covered parties. It may nevertheless produce nationwide practical change if the platform standardizes operations. That distinction between what a coalition can practically change and what it can formally command supplies the framework for Chapter 3’s question: whether practical spillover can compensate for the absence of comprehensive federal legislation.

Chapter 3: Testing the Coalition Model—Meta, Historical Parallels, and Its Limits

3.1: Meta as a Bounded Test of State Regulatory Power

Learning objectives

• Reconstruct the publicly described structure and allegations of the October 2023 multistate action against Meta.
• Separate allegations, reported evidence, procedural developments, defenses, and adjudicated findings.
• Apply all eight stages of the Chapter 2 enforcement pathway to identify what the case illustrates and what remains unresolved.

Jordan reads that dozens of states sued Meta and wonders whether the headline means that Meta’s conduct has already been proven unlawful. The correct starting point is procedural and evidentiary discipline. The action demonstrates substantial coordination among state governments, but a coalition is not a judgment, a complaint is not proof, and a reported association is not automatically legal causation.

%[HALF left](Meta Platforms headquarters in Menlo Park, California._HALF.webp)[Real] %An aerial view of the Meta Headquarters sign and campus in Menlo Park, California.

What the Coalition Alleged and How the Case Was Filed

The supplied Associated Press account reports that, on October 24, 2023, a coalition involving 41 states and the District of Columbia announced coordinated actions against Meta. Thirty-three states filed a federal complaint, while other participating states pursued related actions in state courts. This structure matters because it combines a common institutional effort with multiple legal proceedings. It does not establish that every state asserted identical claims or relied on the same law.

A multistate coalition is an organized group of state governments pursuing related enforcement objectives. Federalism describes the allocation of governing authority between the federal government and the states. Here, the relevant form is sometimes described as uncooperative federalism: states use their own authority to challenge or supplement national policy without formally replacing federal law. #(Bulman-Pozen and Gerken, 2009, Uncooperative Federalism)

The states’ reported allegations concerned three broad subjects:

• Addictive or deceptive product design.
• Misleading safety statements.
• Collection of children’s information.

An allegation is a claim that a party must still support through admissible evidence and applicable legal standards. The wording of a complaint reflects the states’ theory of the case; it does not establish that Meta intentionally harmed children or that any challenged feature legally caused a particular injury.

The Chapter 2 enforcement pathway is most useful here as a diagnostic tool rather than as a prediction of outcome. Applied to this matter, its eight stages can be traced as follows:

flowchart LR
  s1["1 Issue recognition"] --> s2["2 Investigation and record"]
  s2 --> s3["3 Authority selection"]
  s3 --> s4["4 Coalition and jurisdiction"]
  s4 --> s5["5 Federal and state filing"]
  s5 --> s6["6 Motions and discovery"]
  s6 --> s7["7 Settlement or adjudication"]
  s7 --> s8["8 Compliance and spillover"]

The first stage identifies a public problem; the second develops a factual record. The public announcement shows that the states had reached a coordinated enforcement posture, but the supplied materials do not provide the full investigative record, including the precise use of any civil investigative demands or subpoenas. The third stage asks which legal authorities were selected. Possible categories include state unfair or deceptive acts or practices laws, state privacy laws, common-law authority, or parens patriae authority, but the supplied account does not establish that all states relied on these sources or used them in the same way.

The fourth stage concerns coalition formation and jurisdiction, meaning a court’s legal authority to hear a dispute involving particular parties, conduct, or claims. The October 2023 announcement clearly illustrates coordination across jurisdictions. The fifth stage is formal filing: the 33-state federal complaint initiated federal litigation, while related state actions initiated separate proceedings. The sixth stage includes pretrial motions and discovery, the court-supervised exchange of information used to test factual and legal claims. The supplied materials identify these as litigation stages, not as proof that either side has prevailed.

The practical influence of the coalition and its formal limits should be kept separate:

Dimension	Practical influence	Formal limitation
State authority	States can coordinate state police powers and related enforcement resources.	Each state must connect its claims to valid authority and a proper jurisdiction.
Coalition structure	Shared objectives can increase pressure on a national platform.	Participation does not make the states’ claims identical or jointly conclusive.
Litigation	Federal and parallel state filings can develop several legal pathways.	A complaint begins litigation; it does not constitute a judgment.
National effect	A platform may respond uniformly to avoid operating under conflicting demands.	States do not automatically obtain formal nationwide regulatory authority.

Compared with earlier state-led enforcement models, the important continuity is institutional: dispersed governments can use litigation to address a perceived regulatory gap and influence a national firm. The analogy has limits. Digital platforms combine product architecture, data practices, content governance, and speech-related choices, while the claims and procedural postures of participating states may differ. The comparison therefore illuminates coalition strategy, not the merits or outcome of this particular action.

What the Case Does Not Yet Establish

Maya’s evidence-status rule is simple: identify what kind of statement is being made before evaluating its importance. The following classification provides a formative check without converting public reporting into adjudicated fact.

Statement or material	Proper classification
The states claimed that Meta used addictive or deceptive design.	Allegation
Research associated social-media use with some adolescent mental-health outcomes.	Reported research evidence and background, not platform-specific legal proof
Thirty-three states filed a federal complaint and other states pursued related state actions.	Procedural development
Meta could argue that Section 230 or the First Amendment limits the claims.	Defense or legal argument
A court has already found that Meta’s conduct caused legally recognized injuries.	No adjudicated finding is supplied; treating this as established would be an error

The final row is especially important. The supplied materials identify allegations, litigation stages, and unresolved defenses, but no final judicial determination proving the allegations. Media reports and whistleblower disclosures may inform public debate or guide investigation, but they are not substitutes for a tested evidentiary record.

Causation presents a separate problem. Research reviews have reported associations between social-media use and depression, anxiety, or psychological distress among adolescents, while also identifying methodological limitations and uncertainty about the direction of causation. #(Keles, McCrae, and Grealish, 2020, A Systematic Review: The Influence of Social Media on Depression, Anxiety and Psychological Distress in Adolescents) Broader scholarship cautions against treating digital-technology effects as uniform across users or contexts. #(Odgers and Jensen, 2020, Annual Research Review: Adolescent Mental Health in the Digital Age: Facts, Fears, and Future Directions)

A population-level association can be relevant evidence, but it does not by itself prove that a particular platform’s challenged conduct legally caused a recognized injury. Legal causation generally requires a connection among the specific conduct, the mechanism of harm, the timing of the injury, and the applicable legal standard. Researchers and courts must also consider confounding factors, selection effects, reverse causation, and measurement problems. #(Shadish, Cook, and Campbell, 2002, Experimental and Quasi-Experimental Designs for Generalized Causal Inference) #(Pearl, 2009, Causality: Models, Reasoning, and Inference)

To move from a deceptive-design allegation toward a legally sufficient finding, the record would need to address matters such as:

• Which design feature was challenged.
• What representation, omission, or practice allegedly misled users.
• Which users experienced the relevant exposure.
• What legally recognized injury occurred.
• How the feature produced that injury and how alternative explanations were evaluated.

Section 230 and First Amendment defenses add legal-characterization problems. Section 230 is a federal statutory immunity framework that can become relevant when liability is based on a service’s treatment or publication of information supplied by another person. The central question is not simply whether the defendant operates a platform. It is how the claim characterizes the challenged conduct: third-party content, content moderation, editorial activity, product design, or the platform’s own business practice. Scholarly analysis recognizes that this characterization can shape Section 230 disputes. #(Citron and Wittes, 2017, The Internet Will Not Break: Denying Bad Samaritans § 230 Immunity)

The supplied materials do not establish a case-specific ruling on whether Section 230 applies to these claims. The action therefore does not prove either that Section 230 automatically protects all platform-related conduct or that it categorically excludes design-related claims.

The First Amendment defense requires similar precision. Government regulation may raise different constitutional questions when it targets speech, editorial or curation choices, consumer disclosures, privacy practices, or product design. Digital-platform disputes can involve both private platform governance and government regulation, but those categories should not be collapsed. #(Balkin, 2018, Free Speech in the Algorithmic Society: Big Data, Private Governance, and New School Speech Regulation) The relevant analysis would require identifying the exact state requirement, its purpose, the conduct regulated, and the burden imposed. No case-specific First Amendment holding is established by the supplied account.

Other limits remain relevant. Federal preemption occurs when federal law displaces or limits state authority. The Dormant Commerce Clause refers to constitutional limits on state measures that improperly burden or discriminate against interstate commerce. These doctrines, along with jurisdictional requirements and ordinary pleading standards, may constrain state enforcement, but the supplied materials do not resolve their application here.

The seventh and eighth stages—resolution and implementation—also remain unsettled in the supplied record. A settlement is a negotiated resolution rather than a judicial finding after trial. A consent decree is a court-approved resolution that can impose enforceable duties. A compliance obligation is a duty created by a judgment or negotiated instrument; a compliance monitor or structural remedy could oversee or require system-wide changes. No final settlement, consent decree, compliance monitor, structural remedy, or nationwide rule is established by the materials provided. The fixed procedural date is October 24, 2023, and any claim about the current status of later motions, discovery, settlement, dismissal, trial, or appeal requires verification through a current court docket or government source. Information described as current in mid-2024 should not be assumed current now.

Conclusion

The Meta action demonstrates that coordinated state authority can produce substantial practical pressure on a national platform. It shows how coalition formation, federal and parallel state filings, pretrial motions, discovery, and unresolved defenses can connect the eight stages of enforcement. It also illustrates how formal legal authority remains divided even when a platform’s response may have practical effects beyond one state.

The action does not yet establish liability, intentional harm to children, platform-specific legal causation, the inapplicability of Section 230, or the failure of a First Amendment defense. Nor does coalition size create a binding national standard. Its broader significance lies in the tension between adaptive enforcement and fragmentation: litigation can respond to emerging harms through existing state authority, but regulation by litigation may produce inconsistent obligations or major policy changes without a uniform legislative or rulemaking process. Legitimate state enforcement therefore depends on transparent allegations, tested evidence, careful attention to jurisdiction and constitutional limits, and a disciplined distinction between practical spillover and formal command.

3.2: Comparison and Transfer—From Tobacco Settlements to Hypothetical AI Harm

Learning objectives. By the end of this subchapter, you should be able to compare tobacco, mortgage, emissions, and platform enforcement; identify which coalition features transfer across sectors; distinguish state AG coalitions from specialized agencies; analyze policy diffusion; and apply the framework to a hypothetical AI chatbot without treating allegations as proven violations.

A historical timeline tests whether coordinated state enforcement is genuinely new. The hypothetical chatbot then asks whether the same institutional logic survives when the product is generative AI rather than social media. The Meta action discussed earlier is a live and contested application: it demonstrates the possibility of coordinated state pressure around a national platform, but it does not by itself prove wrongdoing, causation, or a general legal rule.

%FULL[infograph] %A timeline comparing tobacco, mortgage, emissions, and technology enforcement by harm, coalition, federal participation, remedy, and practical national effect.

What Historical Comparisons Reveal—and Conceal

Federalism is the division of public authority between state and federal institutions. In some areas, authority is concurrent, meaning that state and federal institutions have overlapping or complementary powers. A multistate coalition coordinates several state actors around investigation, litigation, negotiation, or public pressure. Coordination can affect a national company even when no single state can formally command nationwide conduct.

Case	Harm or risk	Coalition and federal role	Legal model	Remedy and practical spillover	Limitation
Tobacco Master Settlement Agreement	Product risks, public-health consequences, and state Medicaid expenditures	1998 agreement involving 46 states, the District of Columbia, and territories; primarily a coordinated state settlement	Litigation followed by settlement	Advertising restrictions, health-related provisions, and substantial payments; effects extended beyond individual states	Distinctive evidence, public-health history, and fiscal claims do not automatically transfer to technology
National Mortgage Settlement	Borrower harm, servicing failures, information asymmetry, and institutional incentives	2012 hybrid initiative involving 49 states, federal agencies, and major mortgage servicers	State-federal investigation and negotiated settlement	Borrower relief and servicing reforms; national servicers had incentives to standardize practices	It was not a purely state-created model and depended on a particular financial and contractual structure
Environmental and auto-emissions enforcement	Pollution and alleged emissions violations involving technical product systems	California’s Air Resources Board, other state actors, federal Clean Air Act enforcement, and Volkswagen-related settlements	Specialized regulation, agency enforcement, litigation, and settlement	Potential changes in manufacturer practices and product systems	CARB is a specialized regulatory agency, not an AG coalition; federal preemption and technical statutes matter

The transferable lesson is institutional rather than substantive. States can pool information, share costs, present a larger negotiating bloc, and make a national-market actor consider operational changes. A settlement may create a compliance obligation for participating parties without becoming a national statute or an adjudicated finding. This distinction, also central to the settlement-and-spillover analysis in Chapter 2, Subchapter 3, separates formal legal authority from practical spillover.

Tobacco illustrates both the power and the danger of analogy. Its agreement combined litigation, advertising restrictions, health provisions, and payments against a background of extensive product-risk evidence and public expenditures. Its success does not supply a general technology template. Historical analysis emphasizes the distinctive social and public-health record of cigarettes rather than a generic theory that every national product can be governed in the same way #(Brandt, 2007, The Cigarette Century).

The mortgage settlement shows why federal participation must be identified precisely. State AGs, federal agencies, and major servicers operated within a shared initiative. It therefore illustrates cooperative enforcement, not state independence from federal institutions. Its combination of borrower relief and servicing reform also shows how a remedy can address both past injury and future organizational practices #(Aalbers, 2012, Subprime Cities: The Political Economy of Mortgage Markets).

The emissions example adds another warning. California’s Air Resources Board exercised specialized regulatory capacity, while state AGs and federal authorities occupied related but distinct roles. The example demonstrates concurrent authority, not a universal formula for producing one national standard. Environmental federalism can encourage innovation, but it can also produce variation, conflict, and disputes about preemption #(Klyza and Sousa, 2008, American Environmental Policy: Beyond Gridlock).

Policy diffusion is the spread of a law, enforcement strategy, settlement term, or product-design practice. It can occur through:

• Imitation, when policymakers copy another jurisdiction.
• Learning, when policymakers respond to evidence about another jurisdiction’s results.
• Competition, when jurisdictions act to protect residents or influence economic activity.
• Legal compulsion, when a superior authority requires compliance.
• Market adaptation, when a company adopts one operational standard because maintaining several systems is inefficient.

These mechanisms are not interchangeable. A platform may revise one product for all users because separate systems are costly, not because one state created formal national law. Youth-privacy and age-appropriate-design models may diffuse through imitation or market adaptation, but diffusion is not automatic and does not guarantee one settled standard #(Shipan and Volden, 2008, The Mechanisms of Policy Diffusion).

flowchart LR
  n1["State action"] --> n2["Other states imitate or learn"]
  n1 --> n3["Company adopts one system"]
  n1 --> n4["Federal institution responds"]
  n2 --> n5["Similar state policies"]
  n3 --> n6["Practical national spillover"]
  n4 --> n7["Formal federal requirement"]

The comparison check is therefore not whether every case “looks like” tobacco. It is whether the cases share a national-market actor, a lawful institutional tool, a reason to coordinate, evidence supporting a theory of harm, and a remedy capable of implementation. What transfers is the possibility of coordinated pressure. What may not transfer is the evidence, legal authority, federal arrangement, market structure, or product-risk history.

%[HALF left](Archival photograph of a tobacco advertising restriction or _HALF.webp)[Real] %This archival photograph shows public health advocates and officials during a landmark tobacco settlement campaign.

Applying the Model to a Hypothetical AI Platform

Assume a chatbot used by minors allegedly collects sensitive data and provides unsafe mental-health advice. These are hypothetical allegations, not adjudicated findings. The analysis must separate unsafe output, data practices, consumer representations, and causation. Research on language-model risks treats unsafe outputs, privacy risks, bias, and misuse as related but distinct categories #(Weidinger et al., 2022, Taxonomy of Risks Posed by Language Models). Privacy analysis also asks whether collection, retention, and use fit the context and expectations of users, especially minors #(Nissenbaum, 2010, Privacy in Context: Technology, Policy, and the Integrity of Social Life).

The eight-stage enforcement pathway shows how dispersed authority could become coordinated pressure:

1. Perceived harm. Reports, testing, complaints, or incidents could create concern about unsafe advice or sensitive-data practices. Concern initiates inquiry; it does not establish liability.

2. Relevant state authority. State AGs might examine UDAP laws, which address unfair or deceptive acts or practices, state privacy laws, common-law authority, or parens patriae authority where applicable. The scope of each authority would depend on the jurisdiction and facts.

3. Federal role or gap. The FTC could investigate or enforce within its statutory authority, and the DOJ could pursue applicable federal litigation or coordinate with other institutions. If existing federal law does not clearly address a practice, that may constitute a regulatory gap; uncertainty about a gap is not proof that states may fill it without limits.

4. Coalition rationale. States might coordinate to share technical evidence, reduce duplicative demands, and address a platform using common code and national data systems. A coalition can increase practical leverage, but coordination does not cure an unsupported claim.

5. Procedural path. An AG or federal agency might issue a subpoena or a civil investigative demand (CID), a compulsory request for documents, data, testimony, or other information when authorized. Investigation is fact gathering, not adjudication. The matter could proceed to litigation, negotiation, or both. Legislatures and agencies could separately pursue legislation or rulemaking, which are prospective lawmaking processes rather than enforcement of an existing claim.

6. Bargaining incentives. States may seek consumer protection, information, and prospective safeguards. The company may seek to limit costs, uncertainty, discovery, reputational damage, or the risk of inconsistent orders. Those incentives can support a settlement even when the parties dispute the allegations.

7. Likely compliance effects. A settlement might include disclosures, age controls, data-governance changes, safety testing, monitoring, or another structural remedy—a prospective change to organizational or product practices. A consent decree is a settlement embodied in a court order. An assurance of voluntary compliance is a negotiated commitment whose legal effect depends on governing authority and terms. Implementation means carrying out, monitoring, and reviewing the obligation. None automatically establishes that the alleged conduct occurred.

8. Unresolved questions. Evidence would still need to connect conduct to a legal theory and, where required, to injury and causation. The following distinctions prevent analytical shortcuts:

   • Unsafe output is not automatically a deceptive representation or legally actionable injury.
   • Data collection is not automatically an unlawful privacy practice; purpose, notice, consent, retention, use, and disclosure may matter.
   • A consumer-facing promise must be identified before asking whether it was inaccurate or misleading.
   • Evidence of system-wide risk does not necessarily establish causation or a legally cognizable injury in an individual claim.

flowchart TD
  n1["Reported concern"] --> n2["State AG or federal inquiry"]
  n2 --> n3["CID, subpoena, or discovery"]
  n3 --> n4["Legal assessment"]
  n4 --> n5["Negotiation or litigation"]
  n5 --> n6["Settlement or court determination"]
  n6 --> n7["Implementation and monitoring"]
  n8["Legislature"] --> n9["Prospective statute or rule"]
  n9 --> n7
  n10["Private plaintiffs"] --> n5

This institutional map also clarifies allocation. Legislatures create prospective statutory rules. Agencies may issue rules when delegated authority permits. State AGs and federal agencies investigate and enforce within their authority. Courts determine claims, evidence, causation, and judicial remedies, and may enter a consent decree. Private plaintiffs pursue individual or class claims subject to standing, causation, procedure, and other requirements. The institutions may cooperate, but they do not perform identical functions.

Legal limits could shape both litigation and spillover. A platform might raise Section 230 arguments or First Amendment issues when claims or remedies target chatbot output, expression, or editorial functions. State requirements affecting interstate product design could prompt federal preemption or Dormant Commerce Clause arguments. These doctrines are fact-sensitive; the hypothetical supplies too little information to determine their application. Nor does technical risk alone decide whether a legal claim succeeds, particularly in sensitive domains such as mental health #(Bender et al., 2021, On the Dangers of Stochastic Parrots).

A chatbot settlement could produce a de facto national policy if one design is cheaper than maintaining state-by-state systems. That would be practical market influence, not formal national law. Separate state terms, incompatible obligations, or limits on state authority could instead produce fragmentation. Responsive regulation helps explain why an inquiry might escalate from information gathering to negotiation, monitoring, and stronger action, while polycentric governance highlights both the value of multiple institutions and the need for coordination and accountability #(Ayres and Braithwaite, 1992, Responsive Regulation: Transcending the Deregulation Debate) #(Ostrom, 2010, Beyond Markets and States: Polycentric Governance of Complex Economic Systems).

Conclusion

The historical comparisons reveal a recurring model: dispersed state authority can become coordinated pressure when states share evidence, align procedural strategies, and confront a national-market actor. Tobacco shows the force of a large settlement but depends on distinctive public-health evidence and fiscal history. Mortgage enforcement shows a hybrid state-federal arrangement. Emissions enforcement shows specialized state regulation interacting with federal enforcement, not an AG coalition acting alone. Policy diffusion explains how influence may spread through imitation, learning, or market adaptation rather than direct legal compulsion.

Applied to the hypothetical chatbot, the model requires disciplined sequencing: identify the perceived harm, locate authority, account for federal participation or a regulatory gap, investigate, test evidence and causation, select a lawful procedure, negotiate or litigate, and monitor implementation. The Meta action demonstrates the practical possibility of coordinated pressure, but it does not yet prove every allegation or settle the broader law of AI governance. A final comparison check is to classify each proposed outcome as an allegation, evidence, negotiated commitment, adjudicated finding, practical spillover, or formal legal requirement. The application question is whether adaptive state enforcement can remain legitimate without creating fragmentation or regulation by litigation.

3.3: The Conditional Coalition—Legitimacy, Fragmentation, and the Federal Future

Maya Rivera returns to the central question with a harder version: if coalitions can govern nationally in practice, what should determine whether they ought to do so? The previous analysis showed how federalism, state police powers, concurrent authority, civil investigative demands, litigation, and settlements can convert dispersed state authority into coordinated pressure on a national platform. This final step evaluates the costs and legitimacy of that influence. Federalism divides governmental authority between the national government and the states. A multistate coalition is a coordinated group of state attorneys general (AGs), each acting within a particular jurisdiction, or legally defined territory. Their action may have formal legal authority only within those jurisdictions while producing broader practical spillover if a platform adopts one nationwide compliance system.

Learning objectives: By the end of this subchapter, readers should be able to:

• Evaluate the strongest arguments for and against multistate enforcement.
• Explain regulation by litigation, patchwork risk, political incentives, federal preemption, federal floors and ceilings, and policy diffusion.
• Recommend when state coordination should supplement or yield to federal legislation and agency action.

%FULL[infograph] %A balanced scales illustration comparing state experimentation and rapid enforcement on one side with uniformity, accountability, and reduced compliance fragmentation on the other, connected to possible federal floor, ceiling, or cooperative models.

The Coalition’s Democratic and Economic Trade-offs

The eight-stage enforcement pathway provides a disciplined way to test both authority and legitimacy:

flowchart TD
  n1["Identify alleged harm"]
  n2["Locate legal authority"]
  n3["Form coalition"]
  n4["Investigate conduct"]
  n5["Assess evidence and theory"]
  n6["File, negotiate, or litigate"]
  n7["Adjudicate or resolve"]
  n8["Implement and diffuse"]
  n1 -->|"frames"| n2
  n2 -->|"supports"| n3
  n3 -->|"coordinates"| n4
  n4 -->|"develops"| n5
  n5 -->|"informs"| n6
  n6 -->|"leads to"| n7
  n7 -->|"creates obligations"| n8

Applied to the Meta action described in the supplied Associated Press account, the framework does not permit the reader to skip from a public allegation directly to a conclusion about wrongdoing. Stage one identifies the alleged harm. Stage two asks which state statutes, UDAP laws, State privacy laws, common-law authority, or parens patriae theories might apply, and whether each jurisdiction has a sufficient connection to the conduct or affected people. Stage three concerns coalition design: states may coordinate investigative theories, evidence gathering, and litigation strategy without possessing identical laws. Stages four and five involve CIDs, subpoenas, documents, technical information, complaints, and expert analysis. Those materials may support an allegation, but evidence is not itself an adjudicated finding.

At stage six, the AGs may litigate, negotiate, or use the prospect of litigation to obtain a resolution. A settlement is a negotiated resolution; it does not, by itself, establish that the defendant violated the law. A consent decree is a negotiated order approved by a court, while an assurance of voluntary compliance is generally a negotiated commitment under applicable state law and may not involve judicial approval. The legal effect of each instrument varies by jurisdiction. Stage eight concerns implementation: reporting, a compliance monitor, a structural remedy, or product changes may influence conduct beyond the participating states. That is practical national influence, not automatically a federal command.

Status of a claim or obligation	What it means	What it does not establish
Allegation	A government or private party asserts that conduct was unlawful or harmful	That the conduct has been proven
Evidence	Information offered to support or contest an allegation	That a court has accepted the evidence as conclusive
Negotiated commitment	A defendant agrees to specified conduct to resolve or avoid further dispute	That a violation was adjudicated
Consent decree or assurance of voluntary compliance	A formalized negotiated resolution, with legal consequences that depend on the instrument and jurisdiction	A generally binding rule for every state or platform

This distinction is central to the criticism of regulation by litigation: AGs and courts may shape broad social policy through lawsuits and settlements without the public debate, legislative bargaining, and voter accountability associated with legislation. The concern is strongest when a negotiated remedy reaches subjects that require difficult policy choices or imposes long-lasting organizational duties. Yet the counterargument matters. AGs may be enforcing existing statutes against new forms of conduct rather than inventing entirely new law. Litigation can also expose evidence and produce relief when legislatures are divided or federal agencies move slowly. The United States’ broader reliance on adversarial legalism helps explain why litigation can become a significant policymaking mechanism where statutes are broad and regulatory authority is dispersed. #(Kagan, 2001, Adversarial Legalism: The American Way of Law)

The democratic question is therefore not whether every settlement is illegitimate lawmaking. It is whether the legal authority, evidence, procedure, transparency, and remedy are proportionate to the policy being created. A high-profile case may also reflect political incentives. State AG offices may gain electoral, reputational, or institutional benefits from pursuing visible technology cases. Bipartisan participation can provide a partial institutional filter, but it does not remove political or financial incentives. Political motivation alone proves neither bad faith nor legal validity. #(Nolette, 2015, Federalism on Trial: State Attorneys General and National Policymaking in Contemporary America)

Earlier state-led enforcement, including coordinated tobacco litigation and other multistate consumer-protection campaigns, illuminates the coalition model’s capacity to aggregate claims and pressure national businesses. The analogy has limits. Earlier campaigns often involved different products, evidence, statutory settings, and institutional histories. Digital platforms and AI systems can change continuously, making implementation and technical oversight more central. The comparison therefore demonstrates a recurring governance pattern, not that the Meta action has the same facts, legal merits, remedy, or public-health consequences.

Choosing Between State Experimentation and Federal Uniformity

The coalition’s practical success can create the patchwork problem. Conflicting state standards may increase compliance costs and fragment digital markets. A company may need different disclosures, age-assurance systems, data practices, or safety procedures in different jurisdictions. Coordination can reduce some duplication by encouraging one compliance approach, but it cannot guarantee that all state laws or remedies become identical. State variation is not automatically harmful: it can reveal regulatory weaknesses, reflect local preferences, and support policy experimentation. Privacy expectations are also context-dependent; what is appropriate in one social setting may not be appropriate in another. #(Nissenbaum, 2004, Privacy as Contextual Integrity)

Federal preemption determines how much variation may remain. Express preemption occurs when Congress explicitly displaces specified state requirements. Implied preemption is inferred from federal structure or purpose. Conflict preemption may arise when simultaneous compliance is impossible or when state law obstructs federal objectives. Federal regulation touching privacy, consumer protection, or AI does not automatically preempt every state requirement. Preemption may preserve state enforcement, remedies, or regulation of matters outside the federal scheme. It is a question of statutory interpretation and institutional allocation, not an automatic consequence of federal involvement. #(Nelson, 2000, Preemption)

The choice between a federal floor and ceiling makes the policy trade-off explicit:

Federal design	State role	Principal benefit	Principal risk
Federal floor	States may adopt stronger protections and often retain some enforcement authority	Preserves experimentation and permits faster responses to emerging harms	Leaves compliance variation and possible interstate conflict
Federal ceiling	Federal standard limits additional state requirements	Promotes uniformity and lowers fragmentation costs	Can block states from responding to harms or preferences not addressed federally
Cooperative framework	Federal agencies set direction while states implement, investigate, or enforce within defined boundaries	Combines national coordination with local capacity	Overlapping authority may produce uncertainty and unclear accountability

A floor and a ceiling are therefore not interchangeable labels. Each allocates policymaking power differently. The choice also requires decisions about state AG enforcement, private lawsuits, agency rulemaking, judicial review, and how new technology will be addressed. Cooperative federalism often combines federal direction with continuing state implementation rather than separating authority into sealed compartments. #(Bulman-Pozen & Gerken, 2009, Uncooperative Federalism)

Policy diffusion offers an intermediate path. A state rule, investigation, or settlement may spread through learning, imitation, competition, or pressure to avoid falling behind. A platform may adopt one national practice because separate systems are inefficient. Other states may then copy or adapt the approach. This can create national influence without formal legal uniformity. But diffusion is not proof of effectiveness; visible or politically attractive policies may be copied even when conditions differ. #(Shipan & Volden, 2008, The Mechanisms of Policy Diffusion) State experimentation has long been analyzed as a source of policy innovation across the American federal system. #(Walker, 1969, The Diffusion of Innovations among the American States)

The bounded AI scenario shows why these choices will become more difficult. Suppose a chatbot is marketed to minors and designed to simulate human intimacy. State AGs receive evidence suggesting that users may not understand that the system is artificial or that safeguards may be inadequate for vulnerable users. A coalition could apply the eight stages: define the alleged harm; identify UDAP, privacy, discrimination, or other jurisdiction-specific authority; coordinate; issue CIDs or subpoenas; evaluate product design, disclosures, complaints, and technical records; litigate or negotiate; obtain a judgment or negotiated commitment; and monitor implementation.

Existing law is not irrelevant merely because the technology is new. At the same time, consumer-protection categories may not resolve every question about causation, model behavior, age classification, or safety design. AI harms arise from the interaction of models, data, deployment settings, business incentives, users, and affected communities, not from the model in isolation. #(Selbst et al., 2019, Fairness and Abstraction in Sociotechnical Systems)

At least three constraints should shape any response: first, the quality of evidence connecting the chatbot’s design to a legally cognizable harm; second, possible First Amendment, Section 230, federal preemption, and jurisdictional limits, none of which can be resolved abstractly; and third, the risk that differing state remedies will produce fragmented or technically inconsistent obligations. Administrative capacity and democratic accountability add further constraints. A settlement might produce rapid safeguards, but if it becomes a de facto national policy, affected users, parents, developers, and the broader public need meaningful transparency about who decided what and on what evidence.

Conclusion

The Meta action demonstrates the capacity of coordinated state enforcement to transform dispersed formal authority into national practical influence. It does not, on the supplied record, prove that the alleged conduct was unlawful, that any settlement terms were effective, that consumer welfare improved, or that the process established durable precedent. Those conclusions would require the factual record, procedural history, resolution documents, and outcome evidence.

The strongest case for multistate enforcement is adaptive: states can investigate emerging harms, test existing statutes, respond when federal action is delayed, and generate information for later legislation or agency action. The strongest case against it concerns legitimacy, fragmented obligations, uneven expertise, and the possibility that settlement becomes policymaking without legislative accountability. A nationwide settlement may be more consistent and faster than a patchwork of state statutes; state statutes may provide stronger public debate and electoral authorization. Neither form is automatically more legitimate.

For the AI chatbot scenario, the most defensible recommendation is a hybrid. State coordination should address concrete, jurisdictionally supportable harms and preserve experimentation while federal legislation establishes durable definitions, minimum protections, institutional responsibilities, and rules for preemption. Federal agencies should use delegated authority for evidence-based enforcement and, where authorized, rulemaking; implementation should remain transparent and reviewable. State coordination should yield to federal legislation or agency action when divergent remedies create serious conflict, when a national technical standard is essential, or when the proposed obligations exceed credible state authority. It should supplement—not silently replace—those institutions when evidence is developing, harms are locally experienced, and federal law leaves a genuine regulatory gap.

Conclusion

States do not automatically control national platforms. Federalism gives them concurrent authority through police powers, UDAP and privacy laws, common-law claims, and federally authorized enforcement. When federal action is delayed, attorneys general can coordinate without becoming one national regulator: they share evidence, overcome resource disparities, and gain investigative and bargaining leverage. The process moves from a regulatory gap to a supported legal claim, coalition formation, investigation, litigation or settlement, and monitored compliance. Coordination can create broad practical effects, but it does not erase preemption, jurisdiction, remedy, or federalism limits; a settlement is not proof of liability. The key skill is matching authority to actor, evidence to claim, and strategy to risk.

⟬

1. What is the best description of a multistate coalition’s effect on a national technology platform? A. It becomes a new sovereign with authority to enact federal law B. It automatically creates a nationwide judicial precedent C. It can coordinate pressure that leads to nationwide operational changes without creating federal legislation© D. It eliminates the need for each state to rely on its own legal authority A coalition can pool resources and encourage a platform to adopt one nationwide system, creating practical spillover or a de facto national policy. However, it does not become a new sovereign or create federal law.
2. Which institution is primarily responsible for creating prospective federal statutes? A. Congress© B. A state attorney general C. A trial court D. A private plaintiff Congress creates prospective federal laws. Other institutions may enforce, interpret, or seek remedies under existing law, but they do not perform Congress’s legislative function.
3. Which step is necessary to turn the statement “a platform harms teenagers” into a legally cognizable claim? A. Assuming that social harm automatically proves a legal violation B. Identifying specific conduct, a legal theory, supporting evidence, authority, and an available remedy© C. Showing only that the platform operates in multiple states D. Obtaining a nationwide policy from the platform before filing a claim A broad social concern must be narrowed to specific conduct and matched to the elements of a recognized legal theory, along with evidence, standing or enforcement authority, and a legally available remedy.
4. What is the difference between a complaint and an adjudicated finding? A. A complaint is a court’s final conclusion, while an adjudicated finding is an initial allegation B. A complaint requests information, while an adjudicated finding is a subpoena C. A complaint contains allegations, while an adjudicated finding is a court’s conclusion after applying legal standards© D. There is no meaningful difference between them A complaint starts or describes a legal action and contains allegations. An adjudicated finding results only after a court evaluates the evidence and applies the relevant legal standards.
5. Why can a regulatory gap encourage state action without giving states unlimited authority? A. Federal delay removes all constitutional constraints on state governments B. States may use existing legal authority, but they remain subject to preemption, jurisdictional rules, and constitutional limits© C. A regulatory gap automatically authorizes every state attorney general to create new statutes D. Federal delay transfers all federal regulatory power to state courts A regulatory gap reflects a mismatch between rapid change and federal action. It may motivate states to use existing powers, but it does not eliminate legal boundaries.
6. Which governance mechanism generally offers the strongest direct democratic mandate and the potential for broadly applicable legal requirements? A. A private settlement B. A civil investigative demand C. Legislation enacted by a legislature© D. A warning issued by an agency Legislation is enacted through the legislative process and creates generally applicable statutory requirements. Settlements, demands, and warnings are narrower instruments and do not normally create universal law. ⟭