The UK Government’s Mythos AI Tests Help Separate Cybersecurity Threat from Hype
- The UK government’s AI Security Institute (AISI) has evaluated Anthropic’s Mythos Preview model, adding independent public verification to previous reports of its capabilities.
- AISI found that while Mythos is not significantly different in tests of individual cybersecurity-related tasks compared to other recent models, it could potentially set itself apart through its ability to chain these tasks into a multistep series necessary for full system infiltration.
- Since early 2023, AISI has been evaluating various AI models’ performance on Capture the Flag challenges. Mythos Preview now completes more than 85% of these Apprentice-level tasks, marking significant improvement over earlier iterations.
Originally published at arstechnica.com. Curated by AI Maestro.
Stay ahead of AI. Get the most important stories delivered to your inbox — no spam, no noise.