On August 7 (local time), artificial intelligence firm Anthropic revealed that it had refined the biosafety protocols of Claude Fable 5. This enhancement significantly minimizes false interceptions by fine-tuning the classifier. Tests indicate that the frequency of bio-related 'downgrade switches' has plummeted by roughly 85% across platforms post-update, enabling users to leverage Fable 5 more seamlessly for biology-centric tasks. Anthropic emphasized that this upgrade will diminish the false interception rate in contexts such as daily health inquiries, experimental result analyses, and biology education, while also bolstering clinical task support for healthcare professionals. Nevertheless, for requests pertaining to professional biological research and drug development that carry 'dual-use' risks—including virology, toxicology, and molecular design—Fable 5 will continue to switch to a lower-capability model for processing.