Anthropic’s Mythos AI model has recently come into the spotlight after being accessed by unauthorized users, according to a report from Bloomberg. This model is meant to enhance cybersecurity but can be risky if misused. A member of this unauthorized group, who claimed to be a contractor for Anthropic, revealed that they accessed Mythos using a mix of starting information from the contractor and commonly available internet tools.
The Claude Mythos Preview is designed to find and exploit weaknesses in major operating systems and web browsers. Currently, only a few select companies like Nvidia, Google, Amazon, Apple, and Microsoft have formal access through Anthropic’s Project Glasswing. Governments are also interested in this advanced technology. However, Anthropic has no plans to make it publicly available due to fears that it could be weaponized.
In a statement, Anthropic confirmed they are investigating the unauthorized access, which they believe is limited to a third-party vendor’s environment. They stated that so far, there’s no sign that this breach has affected their systems directly.
Interestingly, the breach occurred on April 7, the same day Anthropic announced limited access to Mythos. The group behind this has not been identified but is reportedly part of a Discord channel that looks for details about unreleased AI models. They used knowledge from a recent data breach involving another Anthropic AI model to make educated guesses about Mythos’s location online.
Since gaining access, this group has reportedly used Mythos regularly. They provided screenshots and live demos of their use to Bloomberg, though they claim not to use it for cybersecurity to avoid detection. Other unreleased Anthropic models also fell prey to this group.
This incident highlights growing concerns in cybersecurity, especially around AI systems. A recent survey by the Cybersecurity and Infrastructure Security Agency (CISA) revealed that 73% of organizations fear potential threats posed by AI. Many experts agree that robust cybersecurity measures are crucial as technology becomes increasingly powerful and accessible.
As AI continues to evolve, ensuring its safe and responsible use remains a big priority. The recent events surrounding Anthropic’s Mythos serve as a reminder of the challenges and responsibilities that come with such innovative technologies.
Source link
AI,Anthropic,News,Security,Tech
