Anthropic faces backlash after Claude Fable 5 added stricter cybersecurity request blocking, with critics warning the safeguards could hinder legitimate white-hat defense work as well as malicious activity.
Uniswap founder Hayden Adams criticized a safety routing mechanism in Anthropic’s new Claude Fable 5 model that detects and blocks cybersecurity-related requests, calling the design choice “an extremely bad decision.” Adams argued the system may not reliably distinguish legitimate white-hat defense work from black-hat attack requests. The dispute highlights a broader trade-off between stronger AI safety controls and product usability, with potential implications for user trust and regulatory scrutiny across the AI sector.