washington – The U.S. government is preparing to make a version of Anthropic PBC’s powerful new artificial intelligence model available to key federal agencies. Concerns that this tool could sharply increase cybersecurity risks; According to a memo seen by Bloomberg News.
Gregory Barbascia, federal chief information officer at the White House Office of Management and Budget, told Cabinet Office officials in an email on April 14 that OMB is putting safeguards in place to allow agencies to begin using Mythos, a highly protected AI tool.
The email does not specifically state that various institutions will have access to Mythos, nor does it provide a timeline for when they will have access or how they will use it. It told top technology and cybersecurity leaders to expect more information “in the coming weeks.”
U.S. authorities have previously encouraged the private sector to use Mythos to improve cybersecurity. Bloomberg reports that the Treasury Department is seeking access to Mythos to uncover flaws in its software.
Anthropic makes Mythos available only to a limited group of companies, such as technology and financial companies, and encourages them to use Mythos to assess their own cybersecurity risks. The company restricted the release of Mythos over concerns that hackers could use its features to steal data or disrupt victims’ networks.
Ahead of the limited release of Mythos, Anthropic briefed U.S. government officials on the model’s full capabilities, including offensive and defensive cyber applications, company officials said.
Leaders from Washington to Wall Street are grappling with the possibility that this model could dramatically make it easier for hackers to find ways to break into sensitive industrial and government computer systems.
“We are working closely with model providers, other industry partners, and the intelligence community to ensure appropriate guardrails and safeguards are in place before releasing modified versions of models to government agencies,” Barbascia wrote in an email with the subject line “Mythos Model Access.”
Antropic declined to comment. Neither Anthropic nor the government would say whether federal agencies received early access to Mythos, if any.
The move to release a version of Mythos to government agencies shows the government’s continued interest in Anthropic’s tools despite ongoing public and legal battles between the company and Trump administration officials.
In 2026, the Pentagon declared Anthropic a supply chain threat under powers normally granted to foreign adversaries over disputes over artificial intelligence safeguards. Anthropic argued the move could cost it billions of dollars in revenue, and in March the company won a court order blocking the government’s ban on its use of the technology.
Inside Anthropic, company leaders became concerned that the model could pose a national security risk after testers were able to use Mythos to find the kinds of critical bugs that would normally require the world’s best hackers to find. These concerns led the company to issue a limited release of this model.
This has similarly set off alarm bells in various parts of the US government.
Among officials focused on national defense, the introduction of Mythos has created significant uncertainty about how to assess cybersecurity risks, people familiar with the matter previously told Bloomberg. Equipping individual hackers with this model and similar AI tools is likely to be the same transformation as turning regular soldiers into special forces operators, the official said.
The same day Anthropic announced the existence of Mythos, U.S. Treasury Secretary Scott Bessent and Federal Reserve Chairman Jerome Powell convened Wall Street leaders in Washington and urged them to use the model to find weaknesses in their own systems. bloomberg
