Anthropic Investigates Unauthorized Access to Mythos AI Cybersecurity Tool

Anthropic is investigating claims that unauthorized users accessed its Claude Mythos AI model, a cybersecurity tool considered too powerful for public release.

The access reportedly occurred through a third-party vendor environment rather than a direct hack.

The company confirmed it is looking into a report that a small group gained access to the Mythos model without the usual permissions.

This access was likely due to misuse of existing access rights rather than a classic cyberattack.

There is no evidence that malicious actors have obtained the model or that Anthropic's systems have been compromised.

However, the incident raises concerns about the security of advanced AI models and the ability of companies to prevent unauthorized use.

Mythos has been released selectively to certain technology and financial firms to help identify vulnerabilities in their systems.

These firms are responsible for controlling access to the model, which was reportedly accessed by individuals who already had permission to view Anthropic's AI models through work with a third-party contractor.

Cybersecurity experts have highlighted the risks posed when powerful AI tools are used outside their intended controls, including potential misuse for fraud or cyber abuse.

#Anthropic #Mythos #AIcybersecurity #unauthorizedaccess #thirdpartyvendor #cybersecurity