Initiative overview
The Framework Convention establishes a set of concrete obligations. States parties must document relevant information about AI systems and make it available to affected persons, ensuring that information is sufficient to allow individuals to challenge decisions made through, or substantially based on, an AI system, as well as to challenge the use of the system itself. Parties must also provide an effective possibility to lodge complaints with competent authorities, offer procedural guarantees and safeguards where an AI system significantly impacts the enjoyment of human rights and fundamental freedoms, and ensure that people are notified when they are interacting with an AI system rather than a human being.
Alongside these procedural rights, the Convention sets out risk and impact management requirements. Parties must carry out risk and impact assessments, addressing actual and potential effects on human rights, democracy and the rule of law, and do so in an iterative manner rather than as a one-off exercise. These assessments must feed into sufficient prevention and mitigation measures, and authorities retain the possibility of introducing bans or moratoria, described as "red lines", on certain applications of AI systems.
In terms of scope, the Convention covers the use of AI systems by both public authorities (including private actors acting on their behalf) and private actors. For the private sector specifically, parties have two options for compliance: they can be directly bound by the relevant provisions of the Convention, or they can take alternative measures to achieve the same outcomes while still respecting their international obligations on human rights, democracy and the rule of law. National security activities are excluded from the treaty's scope, though parties must still ensure such activities respect international law and democratic institutions. The same exclusion applies to national defence matters and to AI research and development activities, unless testing of such systems has the potential to interfere with human rights, democracy or the rule of law.
On governance and oversight, implementation is monitored through a dedicated follow-up mechanism: the Conference of the Parties, made up of official representatives of the states that have ratified the Convention. This body assesses the extent to which the Convention's provisions are being implemented, issues findings and recommendations to support compliance, and facilitates cooperation with relevant stakeholders, including through public hearings on aspects of implementation. The drafting process itself reflected a broad multi-stakeholder approach, involving 68 international representatives from civil society, academia and industry, together with several other international organisations, in addition to the 46 Council of Europe member states, observer states, and a number of non-member states.
