Following the adoption of the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law (CETS No. 225) by the Committee of Ministers, the Parliamentary Assembly should continue to follow closely the impact of AI on human rights.
In its Opinion 303 (2024), the Assembly noted that the Framework Convention could have made it clearer that each government should be obliged to inform its citizens of the use of AI systems in administrative procedures leading to binding legal decisions. The Framework Convention sets out an obligation to notify persons interacting with AI systems that they are directly interacting with such systems rather than with a human (Article 15). This applies for instance to AI-enabled chatbots on government websites but does not cover the use of AI by government bodies in public decision-making processes. For instance, in the Netherlands, the Dutch Tax and Customs Administration, in order to create risk profiles of individuals applying for childcare benefits, used algorithms in which “dual nationality” was used as an indicator of potential fraud. Thus, the algorithms led to racial discrimination, as admitted by the Dutch Government itself in 2022.
The Assembly should recommend clearer standards and regulations for the application of AI systems in the context of public and administrative procedures producing effects on the rights and interests of citizens, including with a view to contributing to the standard-setting activities of the Council of Europe.
The Assembly should look into how AI is being used by member States in public decision-making processes and determine best practices, taking into account human rights standards and principles of lawfulness, fairness, transparency, purpose limitation, data minimisation, accuracy, storage limitation, integrity, confidentiality and accountability and ensure that these standards and principles are respected by governments using AI in these processes.