Anthropic's Claude Fable 5 AI is refusing harmless prompts, including simple greetings like 'hello'.
• Conservative safety guardrails may cause false positives in under 5% of sessions
• Users report flagging of terms like 'cancer' as biosecurity risks
• Concerns exist about silent sabotage of AI/ML development answers

#Blockchain #CryptoNews #BinanceSquare #AI #Anthropic