The White House is preparing to give major federal agencies access to Anthropic's closely guarded Mythos artificial intelligence model.
Gregory Barbaccia, federal chief information officer at the Office of Management and Budget (OMB), told officials at cabinet departments in an email on Tuesday that the OMB is putting protections in place to allow agencies to begin using Mythos.
The email stopped short of confirming agencies would receive access, and provided no timeline for deployment, telling cybersecurity and technology chiefs only to expect more information in the coming weeks.
The move represents a significant shift in the relationship between Anthropic, the San Francisco-based AI company, and the Trump administration, which has been locked in a bitter dispute with the firm since late February.
The Pentagon designated Anthropic a supply chain risk in early March, meaning use of its technology was deemed a threat to national security, the first time an American company had received the label.
The dispute centres on the Pentagon's demand for unrestricted access to Anthropic's models, which the company resisted, insisting its technology should not be used for fully autonomous weapons or domestic mass surveillance.
Anthropic has rolled out Mythos only to a select group of companies and organisations so they can assess its cyber capabilities and strengthen their defences.
Civilian agencies including the Departments of Energy and Treasury are responsible for safeguarding critical infrastructure such as the electric grid and financial system, and accessing Mythos would help them identify vulnerabilities and prepare for potential cyberattacks.
On the day Anthropic publicly disclosed Mythos, US Treasury Secretary Scott Bessent and Federal Reserve chair Jerome Powell convened Wall Street leaders in Washington, urging them to use the model to find weaknesses in their own systems.
Related reading
- Treasury and Fed summon bank CEOs for emergency meeting over Anthropic's new AI model
- Anthropic spots 'emotion vectors' inside Claude
- Attack on OpenAI boss Sam Altman reflects a growing backlash against the artificial intelligence industry
Barbaccia wrote in the email that OMB was working with model providers, industry partners and the intelligence community to ensure appropriate safeguards before potentially releasing a modified version of the model to agencies.
While the White House discussions suggest progress with non-military agencies, the standoff with the Pentagon remains unresolved, with Anthropic still barred from defence contracts as litigation continues.
The recap
- OMB installed protections enabling federal agencies to use Mythos
- The change follows months of public conflict with the Pentagon
- CNBC reported the White House laid groundwork for agency access
