Anthropic Loosens Biology Safeguards on AI Model Claude Fable 5
Anthropic has updated the biology safeguards on its AI model, Claude Fable 5, reducing biology-related fallbacks by about 85%. This change allows more everyday biology questions to be answered, such as interpreting lab results and understanding symptoms. Initially, the model blocked almost all biology queries, leading to a high number of false positives. The update aims to improve user experience by allowing more benign requests to pass through. However, the model still restricts dual-use work like virology and toxicology to prevent misuse. Anthropic's decision reflects the ongoing tension in the AI industry between accessibility and safety. 
