Anthropic asks Claude users for AI interview responses that may aid model training

Anthropic has opened a one-week study asking eligible Claude users to complete AI-led interviews about what they want from AI, with the responses available for research, product improvement and, under Anthropic's supplemental privacy terms, model training. The company says public release of full interview transcripts is optional, but its privacy materials warn that even de-linked interviews can carry re-identification risks once shared.