Anthropic is broadening its Cyber Verification Program, granting additional security professionals access to its Claude models. This expanded access allows for more extensive penetration testing, malware analysis, and vulnerability research by reducing safety restrictions. The program has already identified a significant number of vulnerabilities, including over 33,000 critical or high-severity issues. AI
IMPACT This move could accelerate the discovery of vulnerabilities in AI systems by enabling more in-depth security testing.
RANK_REASON This is an expansion of an existing program and increased access to a product for a specific use case, rather than a new product release or major research milestone.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →