A user attempting to add subtitles to an anime video file using Claude Fable 5.2 encountered an unexpected refusal. The AI generated a text block indicating it would not proceed with the task due to potential copyright issues, despite the user's explicit instructions to ignore such reminders. This response suggests a new layer of content moderation or safety protocols within Claude Fable 5.2 that overrides user commands when copyrighted material is detected. AI
IMPACT Indicates potential new safety restrictions in Claude models that may limit functionality even for seemingly benign tasks.
RANK_REASON User reports unexpected behavior of an existing AI model.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →