原文 · 未翻译
White House briefed Anthropic, Google, and OpenAI on plans for a government AI review process
Key Points
The Trump administration is weighing an executive order to require government review of new AI models before release, reversing its earlier deregulation stance.
The shift was triggered by Anthropic's "Mythos" model, which the company withheld from public release over cybersecurity concerns and which the NSA is already using.
Growing bipartisan public concern about AI and the departure of AI czar David Sacks have further driven the policy change.
After a year of deregulation, the White House is now discussing an executive order that could subject new AI models to government review before their release. The trigger appears to be Anthropic's "Mythos" model.
The Trump administration is weighing a fundamental shift in its AI policy. According to a New York Times report, the White House is discussing an executive order that would establish a working group of tech executives and government officials. The group would examine potential oversight procedures, including a formal government review process for new AI models before their public release.
In meetings last week, the White House briefed representatives from Anthropic, Google, and OpenAI on the plans, the Times writes, citing people familiar with the discussions. The British model is reportedly serving as a template, where several agencies assess whether AI systems meet specific safety standards. A White House spokesperson told The Information that the reports were "speculation" and said any policy announcement would come directly from Trump.
Anthropic's "Mythos" as the trigger
The reversal began in April, according to the Times, after Anthropic introduced its Claude Mythos model. According to Anthropic, the system is so powerful at identifying software vulnerabilities that the company declined to release it publicly, warning it could trigger a cybersecurity "reckoning." The NSA is already using Mythos to assess vulnerabilities in US government software.