US government works to give agencies use of Anthropic’s powerful Mythos AI tool

WASHINGTON – The US government is preparing to make a version of Anthropic’s powerful new artificial intelligence model available to major federal agencies amid concerns that the tool could sharply increase cybersecurity risk, according to a memo reviewed by Bloomberg News.

Mr Gregory Barbaccia, federal chief information officer of the White House Office of Management and Budget (OMB), told officials at Cabinet departments in an e-mail on April 14 that OMB is setting up protections that would allow its agencies to begin using the closely guarded AI tool, Mythos.

The e-mail neither says definitively that the various agencies will get access to Mythos, nor does it provide a timeline for when it might come or how they might use it.

It tells top technology and cybersecurity chiefs to expect more information “in the coming weeks”.

US officials have previously urged private sector organisations to use Mythos to improve their cybersecurity. The Treasury Department has reportedly been seeking access to Mythos in order to uncover its own software flaws.

Anthropic has provided Mythos to only a limited group of technology companies, financial firms and others, urging them to use it to assess their cybersecurity risk. The firm limited the release of Mythos amid concerns that hackers could weaponise its capabilities to steal data or sabotage victim networks.

Before its limited release of Mythos, Anthropic briefed senior officials in the US government on its full capabilities, including its offensive and defensive cyber applications, according to a company official.

Leaders from Washington to Wall Street are grappling with the possibility that the model could make it dramatically easier for hackers to find ways to break into sensitive computer systems in industry and government.

“We’re working closely with model providers, other industry partners and the intelligence community to ensure the appropriate safeguards are in place before potentially releasing a modified version of the model to agencies,” Mr Barbaccia wrote in the e-mail, which had the header, Mythos Model Access.

Anthropic declined to comment. Neither Anthropic nor the government said which, if any, federal agencies have received early access to Mythos.

The move to roll a version of Mythos out to the agencies signals the government’s continued interest in Anthropic’s tools despite a public spat and ongoing legal fight between the company and top administration officials.

The Pentagon in 2026 declared Anthropic a supply chain threat, under an authority normally reserved for foreign adversaries, over a dispute about artificial intelligence safeguards.

The company won a court order in March blocking a ban on government use of the technology, after Anthropic argued the move could cost it billions of dollars in lost revenue.

Within Anthropic, company leaders became worried the model could be a national security risk after testers were able to use Mythos to turn up the types of critical bugs that it would normally take the world’s best hackers to uncover. These concerns prompted the company’s limited release of the model.

It has similarly set off alarms in various parts of the US government.

Among officials focused on national defence, the introduction of Mythos has created uncertainty about how to evaluate cybersecurity risk, a person familiar with the matter told Bloomberg. Equipping an individual hacker with the model, or similar AI tools, would likely be equivalent to turning a conventional soldier into a special forces operator, the person said.

On the day Anthropic publicly disclosed Mythos’ existence, US Treasury Secretary Scott Bessent and Federal Reserve chairman Jerome Powell convened in Washington a meeting of Wall Street leaders to urge them to use the model to find weaknesses in their own systems. BLOOMBERG