The White House promotes opening Mythos to federal agencies

On April 17, according to a memorandum, the U.S. government is preparing to provide a version of the artificial intelligence model Mythos developed by Anthropic to major federal agencies while strengthening security protections, in response to concerns that the tool may significantly increase cybersecurity risks. Balbasia, federal chief information officer of the White House Office of Management and Budget (OMB), said in an email sent to cabinet departments on Tuesday that the OMB is establishing protective measures to allow agencies to begin using this highly restricted AI tool in the future. The email did not clearly state whether agencies would have access to Mythos, nor did it provide a specific timetable or how to use it. The email only informed heads of various technology and cybersecurity departments that they expected to receive more information "in the next few weeks." Anthropic currently only opens Mythos to a small number of technology companies and financial institutions, and encourages it to use it to assess its own cybersecurity risks. Concerned that hackers might use the model's capabilities to steal data or damage systems, the company has strictly limited its release scope.