Anthropic has opened a week-long study asking some Claude users what they want from AI, with an unusual condition: participants can choose to publish their full interview for anyone to read. The invitation offers a chance to put personal experience into the public record, but sharing may be difficult to undo.
Anthropic Watch analysis
What happened
The study runs from 29 September to 6 October 2026. It is open to Free, Pro and Max users of Claude and Claude Code whose accounts are at least two weeks old. Anthropic says an AI interviewer will ask about participants’ experiences with AI, what they would like it to change in areas such as work, school, healthcare or government, and what they want from AI companies. Each interview takes about 15 minutes.
Taking part is optional, and participants can end early to permanently delete a partial interview. Afterwards, they can choose whether to publish the complete interview and the country associated with it. Anthropic says it will not publish Claude account details such as a participant’s name or email address. It also warns that details people include could still identify them, and that while Anthropic can remove its copy on request, it cannot remove copies others have saved. The company says it does not plan to edit interviews before publication, though it may exclude material it considers irresponsible to publish.
Why it matters
This is more than a survey with an unusually chatty form. Public interviews could give researchers and policymakers access to first-hand accounts of how people use AI and what they want from it. But they would also make personal accounts available to employers, acquaintances, governments and anyone else who finds them. A name-free interview is not necessarily an anonymous one; a job, location or past project can be a surprisingly efficient breadcrumb trail.
The results will not represent the public at large. All participants use Claude, and people who opt to publish their interviews may differ from those who do not. Anthropic also acknowledges that its questions shape the answers. Those limits matter if the interviews are later used to describe what society thinks about AI.
Our read
The public option could make this research more useful and open to scrutiny, but it asks participants to trade privacy for a lasting public contribution. Anyone considering it should read their full interview carefully before opting in, remove details that could identify themselves or other people, and treat publication as permanent. “We can take down our copy” is not the same as putting the internet’s toothpaste back in the tube.
Anthropic announced the study on 29 September and says the interviews will be published online after the study and its initial analysis are complete.
What to watch
- How many participants choose to make their full interviews public.
- How Anthropic handles interviews it decides are not responsible to publish.
- Whether later findings clearly distinguish Claude users and public-release volunteers from the wider population.
Discussion spark: Should researchers publish full first-person interviews to make AI research more open, or keep them private even when participants opt in?
Sources and evidence
- What Do You Want from AI? – Anthropic (29 September 2026, 16:39 UTC)
Anthropic Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by Anthropic.