US cybersecurity agency gets full access to Anthropic's Mythos Preview
What's the story
The Cybersecurity and Infrastructure Security Agency (CISA) has been granted full access to Anthropic's flagship Mythos Preview model. According to a US official, the access was given about a week ago. However, there are no clear guidelines from the White House Office of the National Cyber Director on how CISA should use this advanced AI technology.
Implementation issues
Concerns among federal tech leaders
The lack of clear guidelines has raised concerns among federal tech leaders. They have privately complained that the Office of the National Cyber Director (ONCD) has not properly briefed them on how to implement or use Mythos Preview for vulnerability scanning. This comes after Anthropic's selective rollout of Mythos Preview to select organizations, including those in industry and other nations under Project Glasswing.
Access timeline
Shift toward AI technologies for cybersecurity
CISA was not part of the initial Mythos rollout, as reported by Axios in April. However, last week, Nextgov/FCW reported that agency access to this advanced AI model was imminent. This development comes after the Trump administration's shift toward AI technologies capable of quickly identifying vulnerabilities across computer networks, an important aspect of future cybersecurity discussions.
Cybersecurity implications
Benefits and risks of Mythos preview
Models like Mythos Preview can help federal agencies identify vulnerabilities faster by analyzing large amounts of software and system data. They can then surface weaknesses and possible attack paths for human defenders to review. However, these models could also be used by cyber operators in the intelligence community and Defense Department to accelerate their offensive hacking operations.