vulnerability research

AI safety guardrails hinder offensive cybersecurity researchers

Strict guardrails implemented by AI companies like OpenAI and Anthropic are limiting legitimate offensive cybersecurity researchers who hunt for unknown vulnerabilities.