On Monday 30 October, the G7 leaders adopted a Code of Conduct on artificial intelligence (AI). The document is intended, on a voluntary basis, to provide eleven points of guidance for organisations developing, deploying and using advanced AI systems, such as foundation models or generative AI.
The document adopted by the G7 countries stresses that “appropriate measures” should be taken “throughout the development of advanced AI systems, including prior to and throughout their deployment and placement on the market”. In addition, the document encourages organisations to “facilitating third-party and user discovery and reporting of issues and vulnerabilities after deployment”. These organisations should also maintain “appropriate” documentation of reported incidents and mitigate identified risks and vulnerabilities.
Organisations should also publicly report on the “advanced AI systems’ capabilities, limitations and domains of appropriate and inappropriate use”, the document states. It should also be easier for organisations developing advanced AI systems to share information and report incidents.
Just as the European co-legislators did with the AI Act - still undergoing interinstitutional negotiations (see EUROPE 13279/18) - the G7 leaders are advocating a risk-based approach.
They believe that privacy policies and mitigation measures should be developed, and that investments should be made in security, including physical security, cyber security and insider threat safeguards “across the AI lifecycle”. “Reliable” mechanisms for authenticating the origin of content should also be introduced.
In addition, in view of the rapid development of AI, the G7 countries plan to review and update the Code of Conduct, “if necessary”, through multi-stakeholder consultations.
The G7’s proposed Code of Conduct draws on the ‘Hiroshima Process’ on AI, established at the G7 summit in May 2023 (see EUROPE 13271/15). While the EU was involved in drawing up the principles of the Code of Conduct on AI, on Monday 30 October the Commission welcomed the adoption of the document.
“The acceleration of AI capabilities also brings new challenges. (...) I’m now pleased to welcome the G7 international Guiding Principles and the voluntary Code of Conduct, which reflect the EU’s values in promoting trustworthy AI. I call on AI developers to sign and implement this Code of Conduct as soon as possible”, commented the President of the European Commission, Ursula von der Leyen.
See the Code of Conduct: https://aeur.eu/f/9bq (Original version in French by Thomas Mangin)