Anthropic’s Mythos throws White House AI strategy into ‘chaos’: Report

Anthropic’s advanced artificial intelligence (AI) model Mythos threw the White House AI strategy into chaos,” The Wall Street Journal reported Thursday.

US Vice President JD Vance was “alarmed” during a call with leaders of major AI companies, according to the report, warning that AI models like Mythos, capable of independently identifying software vulnerabilities, could enable cyberattacks that local governments are unprepared to defend against.

“We all need to work together on this,” Vance told chief executive officers, including OpenAI’s Sam Altman, Dario Amodei of Anthropic, Elon Musk of SpaceX, Sundar Pichai of Google and Satya Nadella of Microsoft, according to the report.

The April call, following a White House briefing that heightened Vance’s concerns about emerging AI capabilities, triggered a “chaotic administration response to Mythos that threatens to increase government oversight of AI and overhaul the administration’s tech agenda,” the report said.

This marked a shift from previous sentiment about “winning the AI race against China” and removing barriers to deploying models, the report noted, adding that White House is now considering an executive order that would establish formal oversight for the most advanced AI models.

Administration officials asked Anthropic to pause broader access to Mythos for operators of critical digital infrastructure. The White House assigned National Cyber Director Sean Cairncross to lead the response, which may include new safety measures and limits on how private firms influence government use of AI tools.