Council of Europe’s CAHAI adopts feasibility study on AI legal framework

The Council of Europe’s (CoE’s) Ad hoc Committee on Artificial Intelligence (CAHAI) has adopted a feasibility study outlining possible options for a CoE legal framework for the design, development, and application of artificial intelligence (AI) based on human rights, democracy, and the rule of law. One option is to modernise existing binding legal instruments (for instance, an additional protocol to the European Convention on Human Rights could be adopted to enshrine new or adapt existing human rights in relation to AI systems; another possibility could be to modernise instruments such as the Budapest Convention on Cybercrime of Convention 108+ on data protection). Another option is the adoption of a new binding legal instrument in the form of a convention or framework convention. While noting that such an instrument would harmonise rules and obligations across states, and thereby also enhance trust in cross-border AI products and services, the study also emphasises that it may be considered premature to attempt to draft a convention containing detailed legal obligations in relation to AI systems. Moreover, even if a new instrument is approved, a potential drawback lies with the process of entering into force of an international treaty: there is no legal obligation for member states to ratify, approve or accept a new convention even if they have voted in favour of its adoption or have signed it. The study also explores non-binding legal instruments such as recommendations or declaration that would consolidate common principles, or guidance documents to increase the understanding of the relationship between the protection of human rights, democracy, rule of law, and AI. Finally, it is pointed out that an appropriate legal framework will likely consist of a combination of binding and non-binding legal instruments that complement each other. A binding instrument, a convention or framework convention, of horizontal character, could consolidate general common principles, while additional binding or non-binding sectoral CoE instruments could address challenges brought by AI systems in specific sectors. The study will be presented in 2021 to the CoE’s Committee of Ministers.