Anthropic confirms investigation into Mythos unauthorized access, stating the source was a third-party contractor

robot
Abstract generation in progress
ME News message: On April 22 (UTC+8), a small batch of users obtained access to Mythos via a private forum. Anthropic’s testing found that Mythos has high-strength computer security capabilities, able to identify thousands of software vulnerabilities; as a result, it has been indefinitely shelved from a public release and opened up in a targeted manner to about 50 companies and institutions that maintain critical infrastructure and software. The list includes Amazon, Google, and the Linux Foundation, with the goal of helping them patch vulnerabilities before hackers do. Last week, Anthropic CEO Dario Amodei met with senior officials from the Trump administration to discuss how to respond to the cybersecurity threats posed by Mythos, and the White House said the meeting was “productive and constructive.” Anthropic is also negotiating with government agencies regarding early access to Mythos. This outreach may help ease tensions between the two sides, as Anthropic is currently in court with the U.S. government in two courts over disputes concerning the Pentagon model’s usage terms. (Source: BlockBeats)
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments