Anthropic Apologizes for Covert Claude AI Censorship, Promises Transparency

Anthropic apologized for silently implementing new content filters in its Claude 3.5 Sonnet AI model, promising greater transparency after user backlash over unexpected censorship.