What the Global AI Treaty Means for U.S. Law and Legal Practice

In September 2024, the Council of Europe formally opened for signature the world’s first legally binding treaty on artificial intelligence: the Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law. Over 50 nations—including the United States, the United Kingdom, Canada, and many European Union member states—have since signed on. The Convention aims to ensure that the rapid development and deployment of AI systems does not come at the expense of human rights, democratic institutions, or the rule of law.

This post explores the Framework Convention’s structure, guiding principles, benefits, risks, and its likely influence on U.S. AI law. It synthesizes perspectives from government officials, legal scholars, industry voices, and civil society groups who are engaged with this international AI regulation.

A Rights-Based Treaty for the AI Era

Unlike earlier soft-law initiatives like the OECD AI Principles or the voluntary codes from the Global Partnership on AI (GPAI), the Council of Europe’s Framework Convention stands out for its legally binding nature. The treaty is technologically neutral and lifecycle-based, meaning it applies not just to specific types of AI like facial recognition or large language models, but to all AI systems from design and development through to deployment, operation, and decommissioning.

Its core legal obligations include mandatory risk and impact assessments when AI systems are likely to significantly affect individuals' rights or democratic processes. Signatories agree to adopt legislative or administrative measures to ensure those risks are mitigated and that oversight mechanisms are in place. While they are allowed some flexibility in implementation—such as fulfilling the obligations through domestic law that achieves equivalent outcomes—the treaty demands more than aspirational commitments. It establishes clear legal expectations that signatories must align AI with core principles of human dignity, equality, transparency, and accountability.

Importantly, the Convention includes exemptions for national security and defense contexts, as well as for AI systems in early research and development that do not interact with the public. While these carve-outs offer flexibility, they have drawn criticism from watchdog groups concerned about potential loopholes.

Embedding Core Democratic Values in AI Development

At the heart of the Convention is a powerful normative statement: AI must serve and be constrained by the same values that undergird democratic societies. The text enumerates foundational principles such as respect for human dignity, non-discrimination, the right to privacy, and access to effective legal remedies. It also encourages transparency, fairness, and accountability in both public and private sector use of AI technologies.

The treaty requires signatory states provide legal mechanisms allowing individuals to challenge decisions made by or influenced by AI systems, particularly when such decisions have legal effects or otherwise significantly impact people's rights. For example, if an algorithm denies someone access to public benefits or employment, the affected person must have a right to understand the basis of that decision and contest it.

The Convention’s institutional structure includes the establishment of a “Conference of the Parties,” a multilateral body responsible for monitoring compliance, facilitating cooperation, and interpreting the treaty. Though it does not impose penalties like those seen in the EU AI Act, it offers a new forum for mutual accountability among democratic states.

Global Harmonization and Policy Impact

The treaty's international uptake signals a critical step toward global policy harmonization. With more than 50 signatories representing major global economies, the Convention fosters regulatory coherence across regions that are developing their own AI laws, such as the EU AI Act, Canada’s Artificial Intelligence and Data Act (AIDA), and California’s AI Accountability Bill (AB 2013).

By anchoring AI governance in universal democratic principles rather than commercial interests or national competitiveness, the treaty seeks to level the playing field. This approach ensures that human rights and the public interest are prioritized over technological expediency or market dominance.

The treaty also sends a clear message to global tech companies: certain ethical and legal standards will apply regardless of where AI systems are developed or deployed. For multinational developers, that creates a predictable policy environment and reduces the risk of conflicting national obligations.

Potential Benefits of the Treaty

Among its most praised features is the Convention’s grounding in internationally recognized human rights law. For civil society organizations and data protection advocates, this rights-centered approach offers a counterweight to the increasing commodification of personal data and the opaque use of AI in decision-making.

The treaty also serves as a floor, not a ceiling. It does not prevent countries from enacting stricter laws or broader bans on high-risk AI systems. Instead, it provides a common framework that states can build upon. Its flexibility ensures that signatories can adopt approaches best suited to their legal and cultural contexts while still honoring shared obligations.

Another advantage is the treaty’s emphasis on public sector accountability. Many of the most controversial uses of AI—such as predictive policing, algorithmic sentencing, and social welfare triaging—originate from government actors. The treaty’s mandates for transparency, oversight, and redress in public sector deployments are a major step toward responsible digital governance.

Key Risks and Criticisms

Despite its strengths, the Convention is not without shortcomings. Some legal scholars and civil liberties groups have expressed concern about its limited enforcement mechanisms. Unlike the EU AI Act, which includes steep penalties for non-compliance, the Council of Europe treaty relies primarily on peer review, public reporting, and political pressure. Critics argue that without concrete penalties, the treaty’s principles risk being symbolic rather than transformative.

Another criticism is the ambiguity of several provisions. Terms like “significant impact” and “appropriate safeguards” are left open to national interpretation, potentially leading to inconsistent application across countries. This lack of precision could hinder the treaty’s effectiveness, particularly when deployed as a shield against harmful practices.

Moreover, the treaty’s exemptions for national security and early R&D activities have raised red flags. There is a risk that these carve-outs could be exploited to avoid scrutiny of high-risk AI applications, particularly in military or surveillance contexts. Some critics also point to the lighter touch applied to corporate uses of AI compared to public-sector deployments.

Finally, the pace of ratification and implementation remains a concern. Although the treaty opened for signature in 2024, it will not enter into force until at least five countries ratify it, including three Council of Europe member states. This process may take years, delaying real-world impact.

U.S. Involvement and Legal Implications

The United States’ decision to sign the treaty in late 2024 was hailed as a significant endorsement of multilateral AI governance. However, signature does not equate to ratification, and the U.S. has yet to translate its commitment into domestic law. That said, the treaty could shape U.S. policy in several important ways. It should be noted that the most recent HR1 signed into law by President Trump specifically declined to impose a proposed federal moratorium on states enacting their own AI laws. While that decision provides a useful testing ground for states to work out what regulations are best suited to AI development while protecting citizens rights, it also ensures for the short term a non-uniform application of AI related laws as well.

The Treaty Offers the U.S. Some Guidance for Future Federal Regulation

First, the treaty offers a framework for federal regulators like the Federal Trade Commission (FTC), Department of Justice (DOJ), and Department of Health and Human Services (HHS) to develop AI oversight mechanisms that align with international standards. Already, the Biden administration’s AI Executive Order and Office of Management and Budget (OMB) guidance echo many treaty principles—such as requiring impact assessments and ensuring transparency in federal AI use.

Second, the treaty could provide a roadmap for future Congressional legislation. Although bipartisan federal AI bills have stalled in recent years, the Convention’s themes—like algorithmic accountability, public notice requirements, and legal recourse for individuals—could serve as the scaffolding for a U.S. AI Act.

Third, it legitimizes and reinforces the actions of individual U.S. states pursuing their own AI regulations. California’s AB 2013, which imposes documentation, bias audit, and notification requirements for automated decision systems, reflects many treaty mandates. As more states adopt such laws, the treaty provides a harmonizing influence that could eventually pressure Congress to enact a national standard.

Comparative Perspectives: CoE Treaty vs. EU AI Act

While both the Framework Convention and the EU AI Act aim to regulate AI in a rights-respecting manner, they differ significantly in approach. The EU AI Act adopts a vertical, risk-tiered regulatory model. It defines specific categories of AI systems—such as "high-risk" or "unacceptable risk"—and imposes detailed technical and legal requirements on developers. It includes heavy fines for noncompliance and is enforceable across the European single market.

The Council of Europe treaty, by contrast, is broader in scope and less prescriptive. It offers horizontal protections grounded in fundamental rights rather than product safety standards. Instead of focusing on particular AI use cases, it emphasizes procedural fairness, public accountability, and democratic governance.

As a result, the treaty is well-suited to serve as a global baseline—one that can support emerging regulatory efforts in countries where specific AI-use classifications may not yet be feasible or politically viable.

The Road Ahead: U.S. Scenarios and Global Leadership

Looking ahead, the United States faces a choice. It can treat the treaty as a symbolic gesture and maintain a fragmented, sector-specific approach to AI regulation. Or it can leverage the treaty as a foundation for enacting a unified, rights-based national AI law.

There are several plausible scenarios. In one, federal agencies harmonize their guidance with treaty principles and gradually build a soft regulatory framework through rulemaking and enforcement. In another, Congress enacts a comprehensive AI rights bill inspired by the Convention, mandating transparency, impact assessments, and the right to challenge AI decisions across sectors.

A third scenario envisions hybrid federal-state regulation, with the federal government adopting a treaty-aligned baseline and states enacting more robust protections where needed. Conversely, a more cynical scenario would see the U.S. sign and shelve the treaty, opting for voluntary industry codes and global competitiveness rhetoric.

Whichever path the U.S. chooses, its leadership will shape global AI governance. A serious implementation of the treaty would signal that the United States is committed not only to AI innovation, but also to ethical stewardship. While there will be many starts to AI regulation, then re-thinking and reformulations, the benefits of international treaties, much like those regulating nuclear weapon proliferation, are obvious. AI is powerful, and not confined to one country. As amazing as it is now in so many realms, this is the worst AI will ever be and its future capabilities are literally beyond our prediction.

Conclusion

The Council of Europe’s Framework Convention on Artificial Intelligence represents a historic effort to ensure that humanity retains control over machines designed to mimic—or even surpass—human cognition. Its legal commitments echo the fundamental values upon which democratic societies are built, and its uptake by dozens of nations reflects a growing consensus among citizens often wary of their own governments that AI must be governed not just by what is possible, but by what is just.

As the world’s leading AI superpower, the United States has a special opportunity to bring these principles home—not only through diplomatic signatures, but through binding domestic law. How the U.S. responds to that challenge will likely be the development of AI (both good and bad) going forward.