Mandated by the Council of Europe Committee of Ministers for the period 2022-2024 to set up an international negotiation process for drawing up a Convention on the development, design and application of artificial intelligence (AI), the Committee on Artificial Intelligence (CAI) held a 3rd plenary meeting on 11-13 January 2023 in Strasbourg during which Chapter V of the “zero” draft Convention was examined and discussed.
The Chair on the Legal and Regulatory implications of Artificial Intelligence has been able to obtain a version of the “zero” draft Convention, dated January 6th, 2023, and shares some insights here as regards this document.
The purpose of the Convention is to ensure that during their lifecycle, AI systems fully comply with human rights, respect the functioning of democracy and observe the rule of law, regardless of whether these activities are undertaken by public or private actors. The design, development and application of AI systems used for purposes related to national defence are expressly excluded from the scope of this Convention. The negotiators seem to agree that such a Convention must be seen first and foremost as a broad framework which might be supplemented by further obligations in more specific fields.
The draft convention first proposes a legal definition of AI systems, which is very interesting since this issue of definition is a major stumbling block in the current discussions on the adoption of the AI Act at the European Union level.
The draft Convention sets obligations for the application of AI systems by public authorities or by any private entity acting on their behalf, as well as for the application of AI systems in the provision of goods, facilities and services by public and private actors. Regarding the latter, the draft convention contains, more specifically, requirements for the preservation of individual freedom, human dignity and autonomy, as well as the preservation of public health and the environment. Interestingly, the Convention directly addresses the issues raised by AI systems with regard to combatting undue influence or manipulation which can have consequences for people’s human rights, including, but not limited to freedom of expression.
Certain fundamental principles such as equality and non-discrimination, privacy and personal data protection, transparency, oversight, safety and safe innovation should serve as guidelines in the design, development and application of artificial intelligence systems. Furthermore, these fundamental principles should be subject to appropriate public discussion and multi-stakeholder consultation in the light of the relevant social, economic, ethical and legal implications. It is thus considered important to promote digital literacy and digital skills for all sectors of the population.
The document also contains measures and safeguards that ensure accountability and redress for all those affected by AI systems, including those with disabilities. Within their jurisdiction and in accordance with their domestic laws, State signatories to the Convention should therefore ensure that those affected by artificial intelligence are able to seek redress for any unlawful harm or damage suffered as regards human rights and fundamental freedoms resulting from the application of AI systems. They should also ensure the availability of procedural safeguards by establishing the right to a human review of decisions taken by an AI system that could affect people’s human rights and fundamental freedoms, by providing the option of interacting with a human in addition to or instead of an AI system. In cases where there is a legitimate public interest, necessary and proportionate restrictions on the exercise of the aforementioned rights may only be provided for by law.
The “zero” draft Convention also encourages State parties to adopt guidelines for conducting impact and risk assessments, which must integrate the perspective of those affected by AI, whose rights may potentially be impacted by the operation of the system. States should also ensure that the adverse impacts resulting from the application of enhanced AI systems are recorded and given due consideration by AI providers and users, and that these are carried out iteratively, at an appropriate rate, throughout the lifecycle of such a system. It should be stressed that these conditions should be considered in relation to the enjoyment of human rights, the functioning of democracy and the observance of the rule of law. Last but not least, the signatory states reserve the right to impose a moratorium or ban on certain applications of artificial intelligence systems where such measures are considered appropriate and necessary by the competent domestic authority.
A 4th plenary meeting of the Council of Europe’s Committee on Artificial Intelligence is to be held 1-3 February and will include an examination of the Preamble and Chapter VII of the revised “Zero Draft”. The Draft [Framework] Convention on Artificial Intelligence, Human Rights, Democracy and The Rule Of Law of the Council of Europe is expected to be finalised by November 15th, 2023.
This Convention follows in the footsteps of the Interinstitutional declaration on digital rights and principles recently adopted by the EU which aims to ensure the protection of people’s fundamental rights in relation to the use of AI.