OECD updates AI principles in response to rapid technological developments

The revised principles of the Organization for Economic Co-operation and Development (OECD) on artificial intelligence (AI) were adopted at the meeting of the Ministerial Council (MCM) held on 2-3. May in Paris.

Taking into account recent advances in AI technologies, including the emergence of generic and generative AI, the revised version of the principles more directly addresses issues specific to AI, particularly those related to the protection of privacy, intellectual property rights, security and information integrity, the OECD said in a statement to the public.

The principles, which now have 47 members, including the EU, and whose general scope guarantees applicability to the development of artificial intelligence worldwide, lay the foundations for the development of action frameworks for risk management and policy-making in this area. The first intergovernmental standard in the field of artificial intelligence, they aim to promote the development of innovative and reliable artificial intelligence, while respecting human rights and democratic values, it is specified.

The OECD’s AI Policy Observatory, which tracks developments since the adoption of the Principles in 2019, shows that venture capital investment in generative AI start-ups has increased ninefold, demand for AI skills has increased by 130%, and the share of large companies using AI have on average, in the OECD area, almost doubled and become more than four times larger than those of smaller structures.

At the same time, AI has received significant attention and efforts from public authorities. As proof, there are more than 1,000 initiatives in around 70 countries and territories.

The OECD says it is becoming imperative to develop and deploy AI systems that can increase productivity, accelerate scientific research, promote environmental sustainability and improve healthcare and education, while respecting human rights and democratic values.

However, risks that particularly burden the protection of privacy, security, equality and well-being are advancing at an unprecedented speed and scale and materialize in the existence of bias and discrimination, the production and dissemination of misinformation and misinformation, and the distortion of public debate and markets.

According to the press release, the main modifications made to ensure that the Principles remain relevant, robust and adapted are specifically intended to address security issues, to mention the growing importance of combating misinformation and disinformation, and to preserve the integrity of information in the context of generative AI. .

With MAP