Policy BriefInternational

Policy brief · Council of Europe

The First International AI Treaty Creates a Rights-Based Framework

The Council of Europe Framework Convention is a legally binding treaty instrument connecting AI with human rights, democracy and the rule of law—but it had not yet entered into force as of 24 July 2026.

Treaty Office status: Framework Convention on Artificial IntelligenceStatus as of 24 July 2026

Read the original source ↗

Conceptual visual

Independent editorial analysis

This is FUURAA’s own editorial analysis of the cited public source, prepared independently from the cited institution. Source materials remain attributable to their authors and publishers; FUURAA is responsible for their selection, synthesis and interpretation. No cited institution has reviewed or endorsed this article unless expressly stated.

External evidence

What the Framework Convention covers

The Council of Europe describes the convention as the first legally binding international treaty in this field. It addresses activities across the AI lifecycle through principles connected to human rights, democracy and the rule of law.

The European Union ratified it on 15 May 2026. As of 24 July, the Treaty Office recorded one ratification and twenty other signatures; the convention had not entered into force because it requires five ratifications, including at least three Council of Europe member states.

FUURAA editorial analysis

FUURAA editorial perspective

Evidence-led analysis in the public interest

An AI civilization needs boundaries expressed in rights and institutions, not only technical safeguards. Efficiency, autonomy and innovation do not override human dignity or lawful accountability.

Global companies should map legal obligations by jurisdiction and also maintain group-level principles that remain meaningful where regulation is incomplete.

Key judgments

The Framework Convention establishes an international legal reference connecting AI lifecycle activities with human rights, democracy and the rule of law.
It is a legally binding treaty instrument, but as of 24 July 2026 it had not entered into force; legal effect depends on entry into force, ratification and domestic implementation.
Rights-based governance complements technical safety by asking who is affected, what institutions are protected and what remedy exists when systems cause or contribute to harm.

The convention places AI within existing public values

The Council of Europe describes the Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law as the first legally binding international treaty in this field. It addresses activities across the AI lifecycle through principles connected to human rights, democracy and the rule of law. This matters because AI governance is not only a question of whether a system performs as specified. It also concerns how power is exercised, whether people are treated with dignity and whether public institutions remain accountable. The convention provides an international reference for these questions without turning one technical design into a universal solution.

Binding does not mean automatically global

The convention opened for signature on 5 September 2024, and the European Union ratified it on 15 May 2026. As of 24 July 2026, the Council of Europe Treaty Office recorded one ratification and twenty other signatures. The convention had not entered into force because it requires five ratifications, including at least three Council of Europe member states. Signature, ratification, entry into force and domestic measures are distinct stages that can produce different obligations and timelines. This precision does not diminish the convention’s importance; it prevents a legally binding treaty instrument from being misreported as a rule already applying to every company, system or jurisdiction.

Lifecycle governance broadens the field of responsibility

A lifecycle approach invites scrutiny from design and procurement through deployment, monitoring and retirement. Risks may arise not only from a model’s output but from the data selected, the institution using it, the purpose assigned, the people excluded and the route for challenging a decision. Technology-neutral and risk-based principles can remain relevant as systems change, but they require institutions to translate broad commitments into operational controls. FUURAA’s editorial view is that impact assessment, accessible information, human responsibility and remedy should be treated as connected practices rather than separate compliance documents.

Rights protection and innovation are not simple opposites

Clear institutional boundaries can increase confidence and make responsible deployment more durable, while badly designed obligations can create cost or uncertainty without improving outcomes. The appropriate question is not whether regulation helps or obstructs innovation in the abstract, but whether a measure is proportionate, understandable and effective for the risk addressed. Global organisations must map formal obligations by jurisdiction and can also maintain group-level principles for places where regulation remains incomplete. Voluntary principles, however, cannot replace applicable law or independent accountability when people’s rights are affected.

Alternative views & uncertainty

What this evidence does not settle

Technology-neutral principles can remain durable across rapid change, but broad language may lead to inconsistent interpretation or provide limited operational guidance.
International alignment can reduce fragmentation, yet different constitutional and legal traditions may legitimately implement common principles in different ways.

Public-interest implications

What this means for different stakeholders

public

People affected by important AI-supported decisions need understandable information, accountable institutions and practical routes to question or challenge outcomes.

organisations / industry

Companies should map lifecycle responsibilities, jurisdictional obligations, impact assessment, oversight, records and remedy rather than relying on product-level safety claims alone.

policy

Implementation should translate rights, democracy and rule-of-law principles into proportionate duties while preserving legal clarity and effective enforcement.

research

Study should examine how jurisdictions implement the convention and whether resulting practices improve rights protection, accountability and public trust.

What to watch next

Whether the convention reaches the five-ratification threshold—including at least three Council of Europe member states—required for entry into force.
How broad principles are converted into impact assessment, oversight, documentation and remedy in real institutions.
Whether compliance activity produces better outcomes for affected people rather than only more procedural documentation.

Conclusion

The Framework Convention gives AI governance a clear institutional reference beyond performance and market adoption. It places human rights, democracy and the rule of law within the lifecycle of AI activity and makes clear that technical capability does not displace public accountability. As of 24 July 2026 it had not entered into force; its influence will depend on further ratifications, jurisdiction-specific legal steps and implementation quality. For organisations, the durable lesson is to build rights and remedy into operations early, while describing legal status precisely and avoiding unsupported claims of universal compliance.

Independence and relevance disclosure

This is FUURAA’s independent editorial analysis of the cited Council of Europe material. It does not imply Council of Europe approval, participation or endorsement and is not legal advice; legal status and obligations must be checked for each jurisdiction.

Forward view

Institutional implications

Lifecycle governance

Consider rights and accountability from design and procurement through deployment and retirement.

Impact assessment

Examine foreseeable effects on people, institutions and democratic processes.

Accessible remedy

People affected by important AI-supported decisions need information and practical routes to challenge outcomes.