Anthropic has briefed the Trump administration on Mythos, its latest AI model deemed too dangerous for public release due to advanced cybersecurity capabilities that could pose significant risks if misused.
The briefing represents a notable shift in the company’s relationship with the federal government, occurring even as Anthropic remains locked in litigation with the Department of Defense. The Pentagon previously designated the AI company as a supply-chain risk following disagreements over military access to Anthropic’s systems for surveillance and autonomous weapons applications. Co-founder Jack Clark dismissed the legal battle as merely a “narrow contracting dispute” while defending the company’s decision to keep government officials informed about breakthrough AI developments.
The administration appears eager to accelerate Mythos adoption beyond government circles. Trump officials have reportedly been pushing major Wall Street firms, including JPMorgan Chase and Goldman Sachs, to begin testing the restricted model. This coordinated outreach suggests the administration views advanced AI capabilities as critical infrastructure requiring immediate integration across key economic sectors, despite the acknowledged security risks.
Clark’s comments also revealed internal disagreement about AI’s economic disruption potential. While CEO Dario Amodei has warned of unemployment levels comparable to the Great Depression, Clark offered a more measured assessment, pointing to current data showing employment weakness primarily among recent graduates in specific industries rather than broad economic displacement.
The Mythos briefings establish a precedent for mandatory government notification of dangerous AI capabilities, potentially reshaping how leading AI companies balance public safety concerns with competitive advantages in an increasingly strategic technology sector.
TECHCRUNCH AI ↗