What do you want from AI?

Anthropic has announced a new study utilizing Anthropic Interviewer to gather user experiences, hopes, and worries regarding artificial intelligence. Running from September 29 to October 6, 2026, the study is open to eligible Free, Pro, and Max users on Claude and Claude Code whose accounts are at least two weeks old. Participants will spend roughly 15 minutes answering questions about their significant moments with AI, desired changes, and expectations for AI developers. A key update in this study allows participants to choose to make their complete interviews and associated countries publicly available, differing from a previous December study where only high-level results and a small number of quotes were published. While Anthropic will exclude Claude account information such as names and email addresses from the published record, the company warns participants that unedited responses could potentially lead to re-identification if sensitive personal details are shared. Interviews made public cannot be fully retracted once saved by others. Participants can stop the process at any time by clicking an end button, which permanently deletes partial records, and they may also choose to keep their interviews private while still allowing Anthropic to analyze them for Societal Impacts research. The gathered data aims to inform Anthropic, other AI labs, and policymakers as they navigate the societal impacts of AI.