The Treaty Framework

The Jurisdictional Arbitrage Problem

The national legislative efforts documented in this series — the GDPR, COPPA, KOSA, the Australian minimum age legislation — share a structural limitation that none of them, individually or collectively, can resolve: the platforms they regulate are global entities whose design decisions are made in one jurisdiction and whose effects are felt in all of them. The EU can fine Meta for GDPR violations committed in Ireland. It cannot change the algorithmic recommendation architecture that Meta deploys globally. Australia can prohibit under-16 accounts in Australia. It cannot require Meta to change the engagement-maximization logic that shapes the experience of the users who are permitted to have accounts. The United States can pass KOSA. The behavioral advertising system that funds engagement-maximized design will continue operating in every jurisdiction where KOSA does not apply.

This is jurisdictional arbitrage: the ability of global entities to configure their operations to minimize the regulatory burden imposed by any individual jurisdiction, while distributing the revenue from those operations across every jurisdiction that lacks the power to regulate them comprehensively. It is not illegal. It is the default operating mode of every global corporation operating in a world of sovereign national regulatory authorities. And it is the foundational structural problem for digital child protection — and for digital sovereignty more broadly — that national legislation cannot solve.

The structure of the arbitrage is specific. Platforms incorporate in jurisdictions with favorable regulatory environments — Delaware for legal purposes, Ireland for EU tax purposes, Singapore for Asian operations. Their terms of service specify arbitration in those jurisdictions. Their content moderation decision-making is centralized in those jurisdictions. National regulators who wish to enforce their domestic law against a global platform must either serve process in the platform's home jurisdiction, negotiate cooperation with the platform's home regulator, or pursue enforcement through the platform's domestic operations — a subset of the platform's global activities that the platform can configure to minimize its legal exposure.

The EU's Digital Services Act and Digital Markets Act represent the most ambitious attempt to address this arbitrage within a single jurisdiction's regulatory framework. By making market access in the EU conditional on compliance with EU standards, the DSA effectively extends EU regulatory reach to global platform operations in ways that purely national legislation cannot achieve. But the EU represents roughly 15% of global internet users. Platform design decisions optimized for the remaining 85% of the global user population — including the US, where platform revenues are highest — are not governed by EU law.

Treaty Precedents That Work

International treaty frameworks have successfully addressed global industries whose operations cross jurisdictional boundaries and whose harms required coordinated multinational response. Three precedents are particularly instructive for understanding what a digital platform treaty framework would need to do and whether it is achievable.

The Chemical Weapons Convention (CWC, 1993)

The Chemical Weapons Convention, adopted by the UN General Assembly in 1992 and entering into force in 1997, prohibits the production, stockpiling, and use of chemical weapons and provides for their destruction. It has 193 states parties — near-universal ratification — and is enforced through the Organisation for the Prohibition of Chemical Weapons (OPCW), which has authority to conduct inspection and verification activities within member states.

The CWC is relevant to the digital platform treaty question not because algorithmic attention systems are analogous to chemical weapons, but because the CWC successfully addressed the core structural challenge that a digital treaty would face: establishing binding international standards for the production of a category of technology, with verification authority vested in an international body rather than in individual states. The CWC's verification mechanism — on-site inspections, data reporting requirements, challenge inspections — is the model for how a treaty body could verify platform compliance with design standards that cannot be assessed from outside the platforms' systems.

The Framework Convention on Tobacco Control (FCTC, 2003)

The World Health Organization's Framework Convention on Tobacco Control, which entered into force in 2005 and now has 182 states parties, is the most directly analogous precedent for a digital platform treaty. The FCTC addressed a global industry — tobacco — whose products caused documented harm, whose marketing practices specifically targeted young people, and whose global production and distribution made purely national regulation insufficient. It did not prohibit tobacco; it established binding standards for advertising restrictions, packaging requirements, taxation, and the protection of minors.

The FCTC's structure — a framework convention establishing general obligations, supplemented by protocols with specific technical standards — is explicitly appropriate for situations where the technical standards are evolving and need to be updated more frequently than treaty text can be renegotiated. A digital platform framework convention could establish the general obligations (design standards, age protection, algorithmic transparency, research access) in binding treaty language, while delegating the specific technical standards to protocols adopted by conference of the parties — the mechanism that has allowed the FCTC to update advertising restriction standards as digital marketing has evolved.

The Paris Agreement (2015)

The Paris Agreement on climate change demonstrates that international consensus on a harm problem does not require consensus on the precise regulatory mechanism. The Agreement's nationally determined contributions (NDCs) framework allows each signatory to specify its own reduction commitments, subject to regular review and a ratchet mechanism that creates pressure for progressive tightening. The Agreement establishes a binding obligation to pursue the 1.5°C target while allowing flexibility in the means of achieving it.

This structure is relevant because one of the obstacles to a digital platform treaty is the disagreement between democratic and authoritarian states about what platform regulation is for. Democratic states want to protect their citizens from harm while preserving freedom of expression; authoritarian states want to control information flows. A framework that mandates outcomes (reduction in documented harm metrics) rather than mechanisms (specific design requirements) might achieve broader participation than a framework that specifies technical standards, while still establishing accountability structures that create pressure for genuine compliance.

Why Digital Platforms Are Harder

The treaty precedents in Section II succeeded. A digital platform treaty faces obstacles that those precedents did not. Understanding the differences is necessary for designing a framework architecture that can be negotiated, ratified, and implemented.

The Free Speech Complication

Chemical weapons have no First Amendment analog. Tobacco advertising has been successfully regulated in most democratic jurisdictions. Platform algorithmic architecture sits at the intersection of commercial activity, information technology, and speech — and different democratic states have different constitutional frameworks for that intersection. The EU's approach (GDPR, DSA, DMA) is broadly regulable under EU law. The US approach is constrained by the First Amendment in ways that are still being negotiated in litigation. A treaty that required the US to implement EU-style algorithmic transparency requirements would face challenges under US constitutional law. A treaty designed to accommodate US First Amendment constraints would be weaker than what EU law already provides.

This is not an insurmountable obstacle. Treaty obligations are implemented through domestic legislation, and domestic legislation is subject to domestic constitutional constraints. The FCTC's advertising restrictions were implemented in the US through existing FTC authority without requiring First Amendment litigation. A treaty framework that targets design practices, monetization structures, and research access — rather than content — has substantially lower First Amendment exposure than a treaty that addresses algorithmic content curation directly.

The Verification Challenge

The CWC's verification mechanism works because chemical weapons are physical objects with physical production facilities that can be inspected. Algorithmic systems are software running on distributed computing infrastructure. Meaningful verification of compliance with algorithmic design standards would require access to platform source code, training data, and behavioral output data — access that no existing international verification body has the technical capacity to evaluate, and that platforms have no incentive to provide voluntarily.

The verification problem is soluble, but solving it requires building the technical infrastructure before the treaty is negotiated. A treaty negotiating process that begins with a multi-year process of developing verification standards — similar to the process that preceded the CWC — would produce a more implementable verification regime than one that negotiates political commitments first and figures out verification later.

The Great Power Complication

The FCTC negotiation was complicated by the tobacco industry but not by geopolitical competition between the world's major powers. A digital platform treaty negotiation would occur in the context of active geopolitical competition between the United States and China over digital infrastructure, AI development, and the global norms governing both. Chinese platforms (TikTok, WeChat, Temu) operate under different regulatory regimes than American platforms (Meta, Google, Amazon). A treaty that imposed design standards developed in the democratic world on globally operating platforms would affect American and Chinese platforms differently, creating a negotiating dynamic that does not exist in the tobacco or chemical weapons contexts.

The US-China complication does not prevent a treaty; most successful international agreements have been negotiated under conditions of great power competition. It does mean that the coalition of states that negotiates the initial framework — likely a coalition of democratic states with significant digital markets — will need to create enough market-access leverage to make participation by non-participating states more costly than compliance.

What International Frameworks Exist

No binding international treaty currently governs algorithmic attention systems, platform design standards, or the protection of minor users from algorithmic harm. The following non-binding frameworks exist and are relevant to understanding what the political landscape for negotiating binding standards looks like.

The table reveals the current international landscape: extensive non-binding guidance, no binding standards, and a significant gap between the sophistication of the problem and the sophistication of the international response. The CRC's general applicability to digital contexts is the most legally significant existing framework, and it has been used by the UN Committee on the Rights of the Child to issue general comment guidance on children's digital rights — but general comment guidance does not create platform obligations.

What a Treaty Would Need to Contain

Drawing on the precedents in Section II, the obstacles in Section III, and the gap documented in Section IV, a treaty framework capable of meaningfully governing global platform design decisions would need to contain five structural elements.

A Binding Age Floor

The treaty's most immediately actionable provision would codify a binding minimum age for social media access across all signatory states, with platform-side liability for verification. This mirrors the Australian model at the international level. The age threshold can be specified in the treaty text (as with the Australian legislation's 16) or delegated to protocols adopted by conference of the parties (as with FCTC's advertising restrictions), depending on the negotiating states' capacity to reach consensus on a specific number. The binding character of the obligation — and the platform-side liability allocation — is more important than the specific threshold.

Design Standards Protocol

A protocol to the framework convention would establish binding design standards for platforms operating in signatory states: default settings requirements for minor users, algorithmic recommendation restrictions, notification timing standards, engagement metric restrictions (prohibiting metrics that are functionally equivalent to addiction metrics — streak counts, social comparison displays, infinite scroll). The protocol mechanism allows these standards to be updated by conference of the parties as technology evolves, without requiring re-ratification of the framework convention.

Research Access and Transparency Requirements

The treaty would require platforms operating in signatory states to provide access to an independent international research body — analogous to the OPCW but with data analysis capacity rather than inspection capacity — to behavioral data necessary for studying platform effects on minor users. Transparency requirements would mandate quarterly publication of algorithmic design parameters and their documented behavioral effects, available to the research body and to national regulators of signatory states.

An Enforcement Body

An international enforcement body, funded by membership dues and equipped with the technical capacity to assess platform compliance with design standards, would be the treaty's institutional core. Unlike the OPCW, which conducts physical inspections, the digital enforcement body would conduct code audits, behavioral data analysis, and incident investigations. Its findings would be referred to national regulators for enforcement within their jurisdictions. The body would have authority to issue compliance determinations and to recommend sanctions — trade measures, market access restrictions — that the conference of parties could endorse against non-compliant states or platforms.

A Liability Minimum

The treaty would establish a minimum civil liability standard for platforms that violate design standards in signatory states — ensuring that national enforcement mechanisms meet a threshold sufficient to deter violation. The Australian AU$50M maximum provides a reference point; the treaty's liability minimum would establish a floor below which national enforcement regimes could not fall, while allowing states to impose higher penalties.

The Pathway to Negotiation

International treaties of the scope described in Section V do not emerge from international consensus; they are negotiated by coalitions of states that create enough collective leverage to make participation by other states more attractive than exclusion. The pathway to a digital platform treaty framework follows the pattern established by the FCTC and CWC negotiations, adapted to the specific political landscape of digital governance in 2026.

Stage 1: Coalition Formation (2–3 years)

The initial coalition would consist of states that already have significant domestic regulatory frameworks for digital platforms — the EU, Australia, the UK, Canada, Japan, South Korea — and that collectively represent a sufficient share of global digital advertising revenue to create meaningful market-access leverage. The EU alone represents approximately 15% of global digital advertising revenue; the full coalition would represent approximately 35–40%. At that scale, platform refusal to comply with treaty standards becomes economically costly in ways that create genuine compliance incentives.

Coalition formation requires negotiating a common position on the core elements: the age floor, the design standards framework, the research access mechanism, and the enforcement body structure. The OECD Recommendation on Children in the Digital Environment provides a starting point for the common position — a document that 38 states have already endorsed as a statement of principle and that could be converted into treaty language through a negotiation process.

Stage 2: Framework Convention Negotiation (2–4 years)

The framework convention — the general obligations document — would be negotiated at a diplomatic conference convened under UN auspices, with participation open to all UN member states. The negotiation would focus on the binding obligations that all states must accept (age floor, design standard commitment, transparency requirements) and the institutional structure (enforcement body, conference of parties, protocol mechanism). Contentious technical questions — the specific age threshold, the specific design standards, the verification architecture — would be deferred to protocol negotiation, maintaining momentum on the framework text.

Stage 3: Protocol Development (Ongoing)

The first protocol would address age verification standards, developed by a technical expert group in parallel with the framework convention negotiation and ready for adoption at the first conference of parties. Subsequent protocols would address design standards (second COP), algorithmic transparency requirements (third COP), and enforcement mechanisms (fourth COP). This staged approach mirrors the FCTC's development process and allows technical standards to be developed by experts rather than negotiated by diplomats.

What the Absence Demands

The absence of a binding international framework for digital platform governance is not a natural condition. It is the result of decisions — to prioritize commercial deregulation, to treat platform growth as a public good, to defer hard governance questions until political pressure made them unavoidable — that were made by identifiable governments at identifiable moments. It is changeable.

The series of papers in this collection — Legal Architecture Papers I through V — has documented the mechanisms of harm (Youth Record), the regulatory frameworks that nominally address them (GDPR, COPPA, DSA/DMA), the legislation that achieved political consensus but not legal force (KOSA), the legislation that did achieve legal force and what it reveals (Australian model), and the international framework that does not yet exist but whose structure is legible from the precedents that do. The documentation is not the work; the documentation is the precondition for the work.

The immediate demand is national legislation designed as treaty preparation. States that enact domestic frameworks should design them with interoperability in mind: age floors that could be harmonized across jurisdictions, design standards that could be adopted into protocol language, research access mechanisms that could feed an international research body. Domestic legislation that is designed as the first tier of a multilevel regulatory architecture is more valuable than domestic legislation designed to operate in isolation.

The intermediate demand is coalition formation. The states with the most sophisticated domestic frameworks and the most significant market leverage — primarily the EU member states, the UK, and Australia — should be moving from unilateral regulation toward coordinated regulatory frameworks. The EU-Australia Dialogue on Digital Regulation, established in 2023, is a beginning. It is not sufficient. A formal coalition with a shared position on treaty elements would be.

The long-term demand is a treaty process. The Framework Convention on Tobacco Control demonstrated that a global industry causing documented harm to a specific vulnerable population — children and adolescents — can be regulated through binding international law, given sufficient political will and organized civil society pressure. The digital platform treaty process requires both. The political will is developing; the civil society pressure is present but not yet organized at the international level in the way that tobacco control advocates organized globally in the 1990s.

The Sovereignty Gap is real. It is also the same kind of problem that the international community has solved before, when the domestic consensus in enough countries was clear enough, and the harm was documented enough, and the alternatives were exhausted enough, that negotiating the hard thing became easier than not negotiating it.

The documentation is now sufficient. The domestic consensus is developing. The alternatives are being exhausted. The hard thing is what remains.

Sources and References

• Chemical Weapons Convention (CWC). Opened for signature January 13, 1993; entered into force April 29, 1997. Organisation for the Prohibition of Chemical Weapons, The Hague.
• WHO Framework Convention on Tobacco Control (FCTC). Adopted May 21, 2003; entered into force February 27, 2005. World Health Organization, Geneva.
• Paris Agreement under the United Nations Framework Convention on Climate Change. Adopted December 12, 2015; entered into force November 4, 2016.
• UN Convention on the Rights of the Child (CRC). Adopted November 20, 1989; entered into force September 2, 1990. United Nations.
• UN Committee on the Rights of the Child. "General Comment No. 25 on Children's Rights in Relation to the Digital Environment." March 2, 2021.
• OECD. "Recommendation on Children in the Digital Environment." OECD/LEGAL/0389. Adopted July 2021.
• European Commission. "Digital Services Act." Regulation (EU) 2022/2065. In force since February 17, 2024 (for all platforms).
• European Commission. "Digital Markets Act." Regulation (EU) 2022/1925.
• Gostin, Lawrence O. Global Health Law. Harvard University Press, 2014. Chapter on Framework Convention on Tobacco Control.
• Floridi, Luciano, et al. "An Ethical Framework for a Good AI Society." Minds and Machines, 2018. On international AI governance.
• Calo, Ryan and Frank Pasquale. "The Scored Society: Due Process for Automated Predictions." Washington Law Review, 2014.
• Office of the Australian eSafety Commissioner. "International Engagement Strategy." 2024.
• UK Department for Science, Innovation and Technology. "Online Safety Act 2023: International Cooperation Framework." 2024.
• Global Partnership on AI (GPAI). "Children and AI: Advancing Child-Centred AI." Working Group Report, 2024.
• Internet Governance Forum (IGF). Policy Network on Meaningful Access, Children Online session. 2024.
• Ruggie, John G. "Business and Human Rights: The Evolving International Agenda." American Journal of International Law, 2007. On international corporate liability frameworks.
• Couldry, Nick and Ulises A. Mejias. The Costs of Connection: How Data Is Colonizing Human Life and Appropriating It for Capitalism. Stanford University Press, 2019. On global dimensions of data extraction.
• Zuboff, Shoshana. The Age of Surveillance Capitalism. PublicAffairs, 2019. On global behavioral data markets.