Anthropic launches second AI attitudes interview study
Anthropic is running a second large-scale study through its Anthropic Interviewer tool, and this time participants can publish their full interviews, which could give researchers and policymakers an open dataset on how people experience AI.
Anthropic has opened a new study that uses its Anthropic Interviewer tool to ask people about their experiences with AI. Participants can choose to make their complete interviews public so anyone can read them, not only Anthropic.
The study asks three questions: what people's most meaningful experiences with AI have been, positive and negative; what they would like AI to help change in areas such as work, school, healthcare or government; and what they want from the companies building AI, including Anthropic.
It follows a study last December in which 81,000 people shared their hopes and worries about AI. Anthropic says those results shaped the Anthropic Institute's agenda, were presented at the World Economic Forum, and still guide its Societal Impacts and Economics research. Only high-level findings and a few quotes from that study were released.
Publishing is optional. If a participant opts in, Anthropic will publish the full, unedited interview and the country linked to the conversation, but not Claude account details such as name or email. The company says researchers could read interviews in their own languages, compare regions and check whether Anthropic's published findings match what participants said.
The FAQ is direct about the privacy risks. Because interviews are not edited, Anthropic says it cannot guarantee participants won't be re-identified, and notes that researchers have already shown AI can piece together identities from small details in Anthropic Interviewer transcripts. It can remove its own copy on request but not copies others have saved, so it advises treating the decision to publish as permanent.
Anthropic has not said when the study closes or when the interviews or results will be released.