The US and China have begun talks on a mechanism for notifying each other of AI safety incidents
The US and China have begun talks on the US China AI Dialogue framework for notifying each other of AI incidents that threaten national security. According to US Treasury Secretary Scott Bessent, the talks follow Trump's visit to Beijing in May.
The US and China have begun talks on establishing a mechanism called US China AI Dialogue, which is intended to allow both countries to notify each other of artificial intelligence safety incidents affecting national security. According to US Treasury Secretary Scott Bessent, the framework should also include further regular meetings between representatives of both countries, at which they should seek to align shared goals and identify threats. Bessent told reporters on Sunday that the shift from "opacity to greater transparency" between the world's two largest AI powers is important. The talks follow US President Donald Trump's visit to Beijing in May.
According to the source, concerns about AI safety have grown in recent months - partly because of an incident in which agents from OpenAI penetrated the Hugging Face system, and because of warnings from employees at Anthropic that AI development could lead to the extinction of humanity. Leaders of leading laboratories - Sam Altman from OpenAI, Dario Amodei from Anthropic and Elon Musk - have called in recent weeks for AI development to slow down until safety standards are established. President Trump has so far largely rejected these calls; his administration argues that the US government has sufficient authority to regulate the industry through existing criminal and regulatory tools. Nvidia chief Jensen Huang takes a similar position, arguing that existing laws on cybersecurity and liability for damages should first be applied to AI companies.
The US is also continuing to restrict exports of advanced chips and chip manufacturing technologies to China as part of its effort to maintain its technological lead in AI. According to Bessent, these export controls were not discussed at the weekend talks. Thomas Wolf, co-founder of the Hugging Face platform, which Nvidia recently announced it would acquire, questioned whether the US can build global cooperation while openly seeking to widen its own lead - he described such an approach as counterproductive.
Why it matters
This is the first documented step towards partial communication between the US and China on AI safety, even though the two countries remain in technological competition - particularly over exports of advanced chips. For companies and institutions following AI policy, it signals that governments are beginning to consider frameworks for addressing cross-border safety incidents, which could affect regulation or reporting obligations in the future. The specific form of the mechanism, its timeline and its enforceability have not yet been described in the sources.
Relevant practical impact
What this means
For a business
The establishment of a dialogue between the US and China on AI safety incidents suggests the beginnings of a framework for cross-border reporting that could, in the future, affect companies operating AI systems with ties to both markets; meanwhile, US restrictions on exports of advanced chips to China continue, affecting companies in the AI hardware supply chain.
Risks and complianceCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.