(Central News Agency, Washington, 5th, comprehensive foreign report) The US Department of Commerce's Center for AI Standards and Innovation today announced that Microsoft, Google, and Elon Musk's xAI have agreed to an accord allowing the US government access to their new artificial intelligence (AI) models for national security risk inspections before their external release.

Reuters reported that the development of advanced AI systems has caused a stir globally in recent weeks due to their potential to significantly enhance hacker capabilities, with US officials and business circles also involved. Anthropic, a US company that launched the Mythos model, has had disputes with the Pentagon over restrictions on the military use of its AI tools.

The Center for AI Standards and Innovation (CAISI) is the main agency for testing AI models in the US government. It stated today that the new agreement reached with Microsoft, Google, and Elon Musk's xAI will enable it to evaluate these companies' AI models before deployment, understanding their capabilities and safety risks.

CAISI Director Chris Fall said in a statement: "Independent and rigorous measurement science is necessary to understand the most advanced AI and its national security implications." (Compiled by Chang Cheng-Chien) 1150505

FACT BOX

  • Source: CNA (Central News Agency)
  • Category: Partnership
  • Organizations: Anthropic