AI Model Achieves 73% Success Rate in Expert-Level Cyber Attacks, Raising Security Concerns
A recent evaluation of Anthropic's Claude Mythos Preview has revealed the AI model's capability to autonomously compromise weakly defended enterprise networks, with a 73% success rate in expert-level capture-the-flag challenges. This milestone marks a significant leap in AI cyber capabilities, outperforming previous models and raising concerns about the security of corporate networks.
The UK's AI Safety Institute tested Anthropic's Claude Mythos Preview for cyber capabilities. For the first time, an AI model autonomously completed a full attack simulation against a corporate network, but the results come with significant caveats. The article Claude Mythos can autonomously compromise weakly defended enterprise networks end-to-end appeared first on The Decoder.