US and China begin talks on AI safety

The US and China are the world's two biggest countries in artificial intelligence (AI). This week, they started talks aimed at making AI technology safer for the whole world.

The question is whether the meetings between President Donald Trump and China's leader Xi Jinping can lead to real agreements. That depends on how much the two rivals actually want to cooperate.

Proposal for a warning system

Trump wants it to look as if the talks are going well. Before Trump and Xi meet on Thursday and Friday, US Treasury Secretary Scott Bessent said the two countries have already discussed setting up a new warning system for AI safety.

If China agrees to participate, it means the two countries will get a fixed communication channel. Here they can warn each other if their AI systems become dangerous or behave unpredictably. According to Bessent, the channel can also be used to share common goals for AI development.

Mistrust could get in the way

But the proposal comes at a time when China may not trust that the US AI safety talks are only about safety, and not about limiting China's options.

Trump has continually tightened export rules, making it harder for China to obtain technology that can be used to develop AI. At the same time, the US Congress is debating new laws to prevent China from gaining access to more advanced computer chips and other hardware.

The situation has become more tense because just two weeks ago the US Department of Justice accused six major Chinese AI companies of stealing technology from American research labs. China strongly denies the accusations.