Anthropic quietly degraded Fable 5 for AI researchers, then apologized
Anthropic shipped a hidden guardrail in Claude Fable 5 that silently degraded responses for developers working on rival AI systems, using methods including prompt modification and steering vectors with no user notification. Within 48 hours of backlash from the AI research community, the company reversed the policy and apologized. The precedent that frontier labs can quietly reshape what developers can build on top of them, however, remains.