Efforts to establish a global system for regulating artificial intelligence face political and technical obstacles, as a clear dispute emerges between the United States and China over the limits of international oversight. The disagreement comes as Western presidents and prime ministers lead moves to draft common rules, while executives at major technology companies remain divided over creating an international body to set standards for developing and using the technology.
The Wall Street Journal reported an exchange of messages between Western leaders as they sought common ground on addressing the risks posed by artificial intelligence. Norwegian Prime Minister Jonas Gahr Støre wrote to his Canadian counterpart, Mark Carney, expressing hope that Canada would join an emerging plan to create a new regulatory framework for the technology. Carney replied: “Of course... we are with you.”
The correspondence coincided with talks between Carney and French President Emmanuel Macron on AI regulation while they were in the French archipelago of Saint Pierre and Miquelon, off Canada’s Newfoundland island. At the same time, Macron’s government was calling on the UN Security Council to prioritise AI and its relationship with security ahead of this week’s UN General Assembly proceedings.
Turning these contacts into practical proposals, Støre and Finnish President Alexander Stubb published a two-page paper stating that artificial intelligence “must remain under human direction, supervision and control”. The paper also proposed establishing a global oversight system to monitor the technology’s development and limit its associated risks.
The discussions form part of efforts led by a group of Western leaders to establish global safeguards for artificial intelligence, an approach that recalls the nuclear arms-control treaties that emerged during the first decades of the atomic age.
US and Chinese backing to determine the international body’s influence
Analysts have warned that any international body responsible for setting standards will have limited influence unless it secures the support of the US and Chinese governments. The distance between the two powers became apparent during a UN Security Council meeting on artificial intelligence held on the sidelines of the UN General Assembly.
Senior officials from Washington and Beijing put forward opposing visions of international AI safety. The US side opposed expanding the scope of regulation, while China showed greater willingness to discuss international controls governing the technology. Michael Kratsios, director of the White House Office of Science and Technology Policy, strongly opposed creating a new global system to control the development of artificial intelligence.
The position puts the US administration at odds with proposals to expand the powers of international oversight institutions, while a number of Western leaders and AI companies seek to establish standards that extend beyond national frameworks.
By contrast, China’s permanent representative to the UN, Fu Cong, called for closer international cooperation on AI regulation. He urged greater coordination between technology development strategies, governance frameworks and technical standards, as well as early work to establish a consensus-based global framework with institutional controls that would help preserve international peace and security.
Fu indirectly criticised the United States, warning that the logic of “we are up against them” serves attempts by “some major powers” to retain their technological advantage through monopoly as they pursue digital dominance.
AI companies propose standards and oppose the body
The statement reflects the widening dispute over which body should set AI rules and how much authority an international oversight system should be given. Alongside government initiatives, executives at major AI companies have put forward their own proposals for regulating the technology and establishing international standards for it.
Dario Amodei, chief executive of Anthropic, proposed a series of measures the Security Council could implement to strengthen cooperation on safety, including an international ban on using artificial intelligence to develop biological weapons. Amodei’s proposals also included creating assessment systems to monitor the development of AI worldwide.
During their addresses to the Security Council, Amodei and OpenAI chief executive Sam Altman called for the creation of a global body to set standards for the technology, an idea the two companies have been promoting for months. The proposal initially received a positive response from some companies and US administration officials before stalling because of opposition from prominent technology executives.
Meta chief executive Mark Zuckerberg, Nvidia chief executive Jensen Huang and SpaceX chief executive Elon Musk told US President Donald Trump that they opposed creating the body.
Also from within the industry, Hugging Face chief executive Clément Delangue called on private companies developing advanced AI models to provide greater transparency.
His position came despite Hugging Face being the target of a widely publicised hack in which AI models that OpenAI had been testing internally were used. Delangue stressed that the AI industry should not cause fear among the public.
