New Delhi: France is convening a high-level meeting at the United Nations Security Council (UNSC) on Wednesday in which OpenAI, Anthropic and Hugging Face CEOs will brief the 15-member council on the risks posed by artificial intelligence.
France holds the rotating presidency of the Security Council for September, coinciding with the opening of the 81st Session of the General Assembly in New York.
A French official said the meeting had been called by Foreign Minister Jean-Noël Barrot because the security implications of artificial intelligence had not received enough attention at the Security Council.
“We consider that artificial intelligence and the threat it could cause has not been sufficiently discussed, especially at the Security Council, and this is why Minister Barrault called for this meeting,” the official said at a media briefing.
French President Emmanuel Macron is expected to host an event focused on protecting children from potential harms associated with artificial intelligence. The initiative brings together governments, technology companies, civil society organisations and foundations, with the stated goal of ensuring that AI benefits children without harming their cognitive and social development.
About 20 countries are expected to participate at various levels, including heads of state and ministers, according to officials.
“The French president will have a major event on the protection of children with AI, which is a coalition bringing together governments, companies, civil society organisations, foundations, with a focus on ensuring that the development of AI benefits our children and do not harm their cognitive and social development, which is a key topic for the future of our society,” the French official added.
The Security Council meeting will address a different set of concerns. “AI has reached a level of development where it raises questions for global security,” the official said. “Now, on several fronts, the malicious use of models, but also the risk of losing control over advanced models, or autonomous weapons, for instance.”
The decision to invite the technology companies directly involved in developing advanced AI systems reflects the growing role of private companies in shaping the technology and its associated risks.
The official added that OpenAI was invited “in light of the recent security incident of this summer,” where independent AI test agents broke out of their sandboxes and accessed external systems, while Anthropic would appear “in light of its proposal on basing the development of advanced AI systems”.
According to the French officials, governments would ultimately have to establish the rules governing the technology and it will be discussed in the UNSC meeting.
The Security Council has previously addressed emerging technologies in the context of international peace and security, but the French initiative puts the rapidly evolving capabilities of generative and advanced AI systems directly before the council.
The first briefer will be Joshua Bengio, co-chair of the International Scientific Panel on Artificial Intelligence, who is expected to offer what the official described as “an independent academic perspective on the threats that AI causes to peace and security”.
He will be followed by Sam Altman, the chief executive of OpenAI, Dario Amodei, the chief executive of Anthropic, and Clément Delangue, the chief executive and co-founder of Hugging Face, the French artificial intelligence company.
Another French official said Paris believed the technology offered substantial benefits but also required international safeguards.
“First, by saying that France has put AI very firmly on the agenda of this UN General Assembly, and that was on purpose because AI is playing an increasingly important role around the world, no need to say about it,” the official added.
“We strongly believe in the immense benefits it could have, but it has to be done well.”
In the UN General Assembly’s 81st session Tuesday, France President Emmanuel Macron had also focused on the risks posed by AI, particularly the fact that most people and countries have little or no say in the debate over how the technology should be regulated.
He argued that decisions about AI should not be left solely to the US, China and a small group of private companies. The stakes, he said, were too high, given that the technology could threaten “the security of our States, of our hospitals, of our banks”.
“We must defend science. We have waited too long to protect our children and adolescents from social media. Everywhere in our societies, we see the consequences,” Macron said.
(Edited by Tony Rai)