Anthropic Investigates Unauthorized Access to Powerful ‘Claude Mythos’ Model
AI startup Anthropic is investigating reports that a group of users gained unauthorized access to its Claude Mythos model, a powerful AI system that was deliberately released to a small circle of trusted companies because of its advanced cybersecurity capabilities.
A report from Bloomberg reveals that Anthropic revealed Tuesday that it was looking into allegations that individuals accessed the model through a system intended for third-party companies doing work on behalf of Anthropic. In a statement, the company said: “We are investigating a report of unauthorized access to Claude Mythos Preview through one of our third-party vendor environments. »
The incident raises serious questions about whether Anthropic, valued at around $380 billion, can effectively protect its most powerful technologies from falling into the hands of bad actors. The company had intentionally limited the release of Claude Mythos Preview to a select group of technology companies, explicitly citing concerns that the model could be misused to launch cyberattacks on a scale and speed beyond human capabilities.
An individual who gained unauthorized access allegedly used their permissions as a contractor for Anthropic to access Mythos. The company said it had no evidence of activity beyond the “vendor environment,” the infrastructure that third parties use to access systems for model development. AI labs frequently employ contractors for responsibilities such as model testing, although Anthropic did not identify which specific vendor was involved in this incident.
This security breach intensifies existing concerns around Mythos, which has already generated significant market turbulence and catalyzed high-level conversations between global financial institutions and regulators. Security specialists have warned that if acquired by hostile parties, the model could allow hackers to identify and exploit software vulnerabilities faster than organizations can deploy patches and fixes.
Anthropic introduced Mythos earlier this month to a carefully selected list of corporate partners, including Amazon, Microsoft, Apple, Cisco and CrowdStrike. The company said these collaborators would have the ability to detect and secure cyber vulnerabilities using Mythos’ sophisticated capabilities before any wider public release.
Breitbart News previously reported that Anthropic suffered a serious security breach when it accidentally leaked aspects of its source code online:
The latest incident comes just days after Fortune revealed that Anthropic had inadvertently made nearly 3,000 internal files publicly available, including a draft blog post describing an upcoming AI model called “Mythos” or “Capybara” that the company says poses serious cybersecurity risks.
This second leak exposed approximately 500,000 lines of code contained in approximately 1,900 files. When contacted for comment, Anthropic acknowledged that “internal source code” had been leaked as part of a “Claude Code release.” A company spokesperson said: “No sensitive customer data or identifiers were involved or exposed. This is a release packaging issue caused by human error, not a security breach. We are deploying measures to prevent this from happening again.”
The instant bestseller Code Red: the left, the right, China and the race to control AIwritten by Breitbart News social media director Wynton Hall, serves as a model for conservatives to create effective policies around AI, not only for the nation, but also for their families. This becomes even more important when AI giants themselves struggle to secure their AI models.
Sen. Marsha Blackburn (R-TN), who was named one of the TIMEThe 100 most influential people in the field of AI, hailed Code Red as a must read. She added: “Few people understand our conservative fight against big tech the way Hall does,” making him “uniquely qualified to examine how to best use the enormous potential of AI, while ensuring that it does not exploit children, creators and conservatives.” Michael Shellenberger, award-winning investigative journalist and founder of Public, calls Code Red “illuminating,” “alarming,” and describes the book as “an essential conversation starter for those hoping to overturn big tech’s autocratic plans before it’s too late.”
Read more about Bloomberg here.
Lucas Nolan is a reporter for Breitbart News covering issues of AI, free speech and online censorship.
