Anthropic Cancels Silent Claude Performance Downgrades, Apologizes After Community Backlash

According to Beating and Anthropic's official statement, Anthropic announced on June 10 that it is cancelling the silent performance degradation mechanism in Claude Fable 5 after facing strong community criticism. The company had previously planned to silently reduce model performance for accounts suspected of training competing AI systems without notifying users.

Anthhropic apologized for the decision and will now use public refusal mechanisms instead. If the system detects attempts to build high-capability AI, it will explicitly deny the request or redirect users to a lower-capability model.