Unauthorized group has gained access to Anthropic’s exclusive cyber tool Mythos, report claims
According to recent reports, a group of unauthorized individuals has allegedly obtained access to Mythos, Anthropic's highly anticipated cybersecurity tool engineered for business protection. The tool, designed to support enterprise security, was intended for a select set of vendors to prevent misuse.
Entry via Third-Party Vendor
The breach reportedly took place through a third-party contractor working with Anthropic. Members of the group accessed Mythos by leveraging information about where Anthropic typically deploys its AI models, as described in a Bloomberg report. Anthropic has confirmed it is investigating the incident, stating that there is currently no evidence the unauthorized activity has harmed its own systems.
Investigation and Community Involvement
Reports suggest that members of a Discord channel, dedicated to uncovering details on unreleased AI models, have been using Mythos and even presented supporting evidence. The person leading the group worked for a contractor with access privileges, and the breach reportedly occurred the same day Mythos was publicly launched.
The group claims their motivation is to experiment with new technology, not to engage in malicious activity. Even so, the unauthorized use of this advanced AI tool highlights risks for enterprise security, as Mythos could potentially be weaponized if it fell into the wrong hands.
Background on Mythos and Distribution
Mythos was released as part of a project known as Project Glasswing, with partners including major companies like Apple. The controlled rollout was meant to prevent malicious actors from obtaining the technology, reinforcing the sensitivity of unauthorized access incidents like this one.
For additional details, you can read the original article by Lucas Ropek on TechCrunch.
