May 13, 2024 - Technology

U.S. kicks off AI safety talks with China

Illustration of a pile of jenga blocks with the colors and stars of the U.S. and Chinese flags

Illustration: Sarah Grillo/Axios

The United States is heading to Geneva this week to start a series of diplomatic talks with the Chinese government about artificial intelligence safety and risk standards.

Why it matters: The U.S. and China are in tight competition to dominate the AI market, both in the private sector and within their own governments.

  • However, the two world powers have yet to agree on what it means to safely use the technologies they're developing.

Zoom in: The United States and China will meet in Switzerland on Tuesday, senior administration officials told reporters during a briefing Friday.

  • Officials from the White House and State Department will lead the U.S. delegation in the talks, while China will bring a delegation co-led by its Ministry of Foreign Affairs and National Development and Reform Commission.
  • The talks will primarily focus on AI risk and safety "with an emphasis on advanced systems," one official said.
  • Officials from the U.S. and China also plan to discuss the work they're doing in their own countries domestically to address AI risks.

Yes, but: The two parties are only planning to exchange initial thoughts on the technical risks posed by AI technologies and do not expect to provide a joint statement or any other "deliverables," the official said.

  • The official also said that the talks are not intended to result in any technical collaboration or cooperation on frontier research.
  • The U.S. government's "technology protection policies" are also not up for negotiation, the official added.

The intrigue: Issues such as how AI can further fuel China-linked election interference are not officially on the agenda, but they could come up, another official said.

What they're saying: "We certainly don't see eye-to-eye with the PRC on many AI topics and applications, but we believe that the communication on critical AI risks can make the world safer," another administration official said.

Catch up quick: President Joe Biden and Chinese President Xi Jinping agreed to start conversations on AI safety and risk during their meeting at the APEC summit in November.

  • The U.S. also landed China as a co-sponsor on a recent AI resolution in the U.N. General Assembly in March.
  • The U.S. is reportedly looking at placing restrictions on China, Russia and other adversarial nations' access to the most advanced models.

The big picture: The safety standards that the U.S. and China follow will likely influence other countries around the world.

  • Both governments have been exploring the ways they can deploy AI technologies throughout their governments and whether their militaries can adopt AI.
Go deeper