Although the EU AI Act has set clear rules for the regulation of AI systems within the EU, the new AI Convention has now created a common framework for AI systems applicable to the US, the UK, the EU and other countries.

The first of its kind, the Council of Europe Framework Convention on AI and Human Rights, Democracy, and the Rule of Law was signed last week by the EU, the UK, the USA, Israel, Norway, Andorra, Georgia, Iceland, the Republic of Moldova and San Marino.

The treaty is intended to create a legal framework covering the entire lifecycle of AI systems and to promote AI progress and innovation, while managing associated risks. 

The AI Convention creates requirements for the signatories to ensure that activities within the lifecycle of AI comply with the provisions of the AI Convention when undertaken by both public authorities and the private sector. However, it should be noted that the AI Convention provides a certain margin of flexibility as to how the provisions of the AI Convention are effected, in view of the underlying diversity of legal systems among the signatories and the wide variety of contexts in which AI might be used. Signatories could therefore, for example, decide to keep making use of existing regulation or improve it or consider adopting new measures, including non-legally binding measures and interpretative guidance.

Although the AI Convention creates a principles-based framework, introducing foundations for AI systems rather than prescriptive compliance measures, the AI Convention provides a clear steer on requirements particularly around the need to display clear information on the logic used in any AI generated decisions and recommendations and the need for technical standards to allow AI systems to develop as trustworthy and scalable tools.

Easily understood AI systems – setting out the factors taken into account and the rationale for AI generated decisions and recommendations: In connection with the AI Convention and the transparency principle, the Council of Europe, in its explanatory report underlines two important aspects of transparency: explainability and interpretability. The term “explainability” refers to the capacity to provide sufficiently understandable explanations about why an AI system provides information, produces predictions, content, recommendations or decisions, which is particularly crucial in sensitive domains such as healthcare, finance, immigration and criminal justice, where understanding the reasoning behind decisions produced or assisted by an AI system is essential. In such cases transparency could, for instance, take the form of a list of factors which the AI system takes into consideration when informing or making a decision.

“Interpretability” refers to the ability to understand how an AI system makes its predictions or decisions and making outputs of AI systems accessible and understandable to experts and non-experts alike. It involves making the internal workings, logic, and decision-making processes of AI systems understandable and accessible to users.

Reliable AI systems – creating AI systems that are trustworthy and scalable through specific technical standards: This principle highlights the importance of establishing measures that seek to assure the reliability of AI systems through key aspects of functioning such as data integrity and cybersecurity. Technical standards can help deliver easily understood and scalable AI assurance and compliance, while encouraging consistency with already existing legal requirements.

As the first legally binding international treaty on AI, the AI Convention is hoped to provide a global standard for managing technological risks in line with shared values. Further, the AI Convention reaffirms the commitment of the signatories to protect human rights, democracy and the rule of law and recalls international legal instruments and treaties of the Council of Europe and the United Nations which directly deal with topics within the scope of the AI Convention.

The treaty will enter into force three months after five signatories, including at least three Council of Europe member states, have ratified it. In addition to the counties that have already signed the AI Convention, countries from all over the world will be eligible to join it and commit to complying with its provisions.

Please also see our previous post on the EU and US position on the scope and content of the AI Convention, Heavyweight Fight, Did the US or EU KO the AI Treaty?