4. The Assembly has always considered that private actors should fall within the scope of such a legallybinding instrument. In its Resolution 2341 (2020), it expressed the view that the instrument should containprovisions to limit the risks of the use of AI-based technologies by State and private actors to exercise controlover people, and that the activity of private actors should be subject to democratic oversight.

5. The Framework Convention, once adopted, will become the first ever international treaty on AI. It isbased on the Council of Europe’s standards on human rights, democracy and the rule of law, which are alsoshared by the non-member States that participated in the negotiations. This is an example of the Council ofEurope’s leadership in developing standards in emerging areas, including the digital sphere, in line with theReykjavik Declaration adopted by the Heads of State and Government in May 2023. Part of the FrameworkConvention’s added value will be its global reach, since it will bring together States from all over the worldwishing to address the global challenges posed by AI with a human rights-based approach. The Assemblytherefore understands that the drafting process has had to accommodate diverse legal and political traditionsand systems, with the result that the draft text often contains very general and abstract provisions, allowing fora certain level of flexibility in its implementation. Its “framework” nature also means that it will need to besupplemented by other binding or non-binding instruments concerning the use of AI in specific sectors ordeveloping certain provisions of the convention further. The Assembly is ready to contribute to the preparationof such instruments.

6. The Assembly is satisfied that most of the key ethical principles endorsed in its 2020 reports arereflected in different provisions of the draft Framework Convention, although some of these principles couldhave been formulated as positive individual rights rather than general principles (for instance, privacy, equalityand non-discrimination). Furthermore, it could have been made even clearer that each individual governmentshould be obliged to inform its citizens of the use of AI systems in administrative processes leading to bindinglegal decisions. Another significant added value of this draft Framework Convention is that it is intended toprotect not only human rights but also democratic processes and the rule of law in the context of AI. AItechnologies have a potential to disrupt the functioning of democratic institutions and processes, for instancethrough interference in electoral processes, disinformation and manipulation of public opinion. They can alsohave an impact on the functioning of the rule of law, including the independence and impartiality of thejudiciary and access to justice. In this regard, the Assembly considers that the interpretation of “democraticinstitutions and processes” and “the rule of law” within the meaning of the draft Framework Convention shouldbe guided by the relevant standards developed over the years by Council of Europe bodies such as theEuropean Court of Human Rights and the European Commission for Democracy through Law (VeniceCommission), as well as by the Reykjavik Principles for Democracy. The drafters however missed theopportunity to cover more specifically the positive uses of AI for democratic processes, for instance improvinggovernment accountability and facilitating democratic action and participation.

7. The Assembly regrets that the draft Framework Convention does not cover to an equal extent publicand private actors. Rather, it introduces a system where each Party will be able to determine in a declarationhow it intends to address the risks and impacts arising from the use of AI by private actors. This is far fromideal for legal certainty and predictability of the obligations imposed by the Framework Convention and is notin line with the positions previously expressed by the Assembly, the Council of Europe Commissioner forHuman Rights and the CAHAI. It also goes against the principle that States have positive obligations toprotect individuals against human rights abuses by private actors, in accordance with the case law of theEuropean Court of Human Rights, the United Nations Guiding Principles on Business and Human Rights andrelevant recommendations of the Committee of Ministers of the Council of Europe. Many AI systems aredeveloped and deployed by private entities, and introducing a differentiated approach for the private sectorcreates a significant loophole.

8. The Assembly therefore strongly calls on all member States of the Council of Europe, when ratifying theFramework Convention and submitting their declarations under Article 3.1 (b), to recognise the full applicabilityof the principles and obligations set forth therein (Chapters II to VI) to activities of private actors, and to reportaccordingly to the future Conference of the Parties under Article 24. It further invites the Conference of theParties to fully use its powers and conduct a proper review of how all Parties comply with Article 3.1 (b). TheAssembly believes that a dynamic interpretation of this provision by the follow-up mechanism set up by theFramework Convention will foster advances over time, through reporting requirements and peer pressure,including with respect to non-member States that may choose not to apply the Framework Conventionobligations to private actors.

II SÉRIE-D — NÚMERO 4 ____________________________________________________________________________________________________________

28