Anthropic Launches Public Interview Study to Gather User Perspectives on AI

Anthropic Launches Public Interview Study to Gather User Perspectives on AI Anthropic has launched a study using its AI-driven Anthropic Interviewer to collect user experiences and perspectives on artificial intelligence, offering participants the option to release their complete interview transcripts to the public. The 15-minute study runs from September 29 to October 6, 2026, for Claude and Claude Code users whose accounts are at least two weeks old. Participants can choose whether to make their full, unedited interviews and associated country public, with Claude account details excluded. Anthropic warns that public transcripts carry re-identification risks and should be treated as permanent, as external parties may save copies. Building on an earlier study of 81,000 respondents, the project creates an open dataset for external researchers despite not being a fully representative sample. Study Scope and Precedent The study runs from September 29 to October 6, 2026, and is open to Free, Pro, and Max users on Claude and Claude Code with accounts at least two weeks old. Each roughly 15-minute session is conducted by the Anthropic Interviewer system, which asks users about significant positive and negative moments with AI, changes they desire in the world, and expectations for AI developers. The project follows a similar survey conducted the previous December involving 81,000 people, which informed the Anthropic Institute's agenda and Societal Impacts research. Unlike the prior iteration, which released only high-level findings and limited quotes, the new initiative allows individual interviews to be shared publicly. Public Release and Privacy Considerations Publishing interview transcripts is optional. Anthropic will exclude Claude account information like names and email addresses, releasing only the conversation and the associated country. However, Anthropic does not edit transcripts and cautions that specific details regarding employers, past projects, or locations could lead to re-identification, noting researchers have demonstrated this is possible using AI. Because outside parties can save published text, Anthropic cannot delete external copies if a user later requests removal. Users can exit early using the 'End interview' button, which permanently deletes partial responses and prevents both publication and internal analysis. Research Aims and Methodological Limitations Anthropic intends for the public transcripts to provide an accessible historical record that policymakers, journalists, and independent researchers can analyze without relying solely on Anthropic's own conclusions. At the same time, the company acknowledges methodological constraints: the study is restricted to existing Claude users and further filtered by those who choose to make their responses public, meaning the resulting data is not representative of the broader public.

Anthropic启动新一轮AI公众调研，允许参与者自主选择公开访谈记录 Anthropic宣布通过“Anthropic Interviewer”面向Claude用户开展新一轮调查，探讨公众对AI的实际体验与未来诉求，并首次允许受访者自主决定是否将完整对话记录向社会公开。 调研于2026年9月29日至10月6日进行，面向注册满两周的Claude及Claude Code用户，耗时约15分钟。 项目继承了去年12月对8.1万人开展的前期调研，旨在将公众观点公开留存，供全球研究者和政策制定者参考。 公开访谈属于自愿选项，虽不包含账户信息但不会经过删改，Anthropic提示参与者存在被重新识别的隐私风险。 受访者可随时中止并永久删除记录，Anthropic也坦承该调研仅覆盖Claude用户群，并非公众的代表性样本。 调研目的与公开机制 Anthropic认为平衡前沿AI的效益与滥用风险不能仅靠AI厂商自主决定。本次调研在前序8.1万人调研的基础上扩展，探讨受访者与AI交往的重大经历、期望AI改变世界的方向以及对AI公司的诉求。与以往仅公布高层结论不同，本次受访者可选择将完整记录与关联国家公开，供全球研究者、决策者和公众审查，成为记录当前AI发展阶段的公共档案。 隐私提示与知情同意 公开访谈属于完全自愿行为。Anthropic不会附带受访者的姓名和邮箱，但公开版本不会经过文本编辑，结合工作经历、城市等零散细节仍有可能被第三方或AI工具重新识别。此外，一旦公开发布，数据无法被完全彻底撤回。系统在访谈过程中会展示敏感信息防范示例，并在结束时让用户通读全文并最终确认，受访者也可在任意时刻点击结束访谈以彻底删除记录。 参与规则与样本局限 该研究在2026年9月29日至10月6日期间运行，适用于注册至少两周的Claude与Claude Code各级别用户，由AI访谈员主导交流。Anthropic同时指出了本研究的局限性：所有参与者均为Claude现有使用者，不能代表广泛的社会公众，愿意公开发言的群体也可能存在选择偏差，加之提问方式也会对回答产生引导效应。

*We’re launching a new study using [Anthropic Interviewer](https://www.anthropic.com/about-anthropic-interviewer) to learn from your experiences with AI, and we’d like you to participate. After you finish, you can decide to make your interview public, so that anyone, not just Anthropic, can read and learn from it. You can [participate here](https://claude.ai/redirect/website.v1.ad057a73-2de4-4b41-8cd9-83fb8a2c83fa/anthropic-interviewer/your-thoughts-on-ai?from=blog).* We are at a pivotal moment in the development of AI, as its growing capabilities mean it becomes potentially more useful and more dangerous. Frontier AI is rapidly accelerating discoveries in [science](https://www.anthropic.com/research/Claude-accelerates-protein-design) and [medicine](https://openai.com/index/diagnose-rare-childhood-diseases/), while at the same time the cost of its misuse grows [more consequential](https://darioamodei.com/post/we-must-pace-the-frontier)—[a single security incident](https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/#core-takeaways-about-this-incident) can cause far more damage than it did a year ago. How to weigh these benefits and risks shouldn’t be left to AI companies alone. Your voice can help guide us, other AI labs, and policymakers as we navigate this uncertain future together. Our big questions are: This project builds on a similar study conducted last December, in which [81,000 people](https://www.anthropic.com/features/81k-interviews) told us their hopes and worries about AI. That study shaped [the Anthropic Institute’s agenda](https://www.anthropic.com/research/anthropic-institute-agenda), was presented at the World Economic Forum to international leaders and decision-makers, and continues to guide our ongoing [Societal Impacts](https://www.anthropic.com/research/team/societal-impacts) and [Economics research](https://www.anthropic.com/research/81k-economics). AI and Anthropic have changed a lot since then. AI has improved, more people use Claude, and people’s feelings toward the technology have shifted. For the previous study, only the high-level results and a small number of quotes were made public. This time, we want your thoughts to be made widely available so that anyone can learn from them. We’ve written a FAQ so you can make an informed decision about whether to make your interview public and what to share. You can choose to [participate here](https://claude.ai/redirect/website.v1.ad057a73-2de4-4b41-8cd9-83fb8a2c83fa/anthropic-interviewer/your-thoughts-on-ai?from=blog). The interview asks about your experiences with AI and what you want from it. We’ll ask about the moments with AI that have mattered most to you, both positive and negative, whether there’s anything about how the world works that you’d like AI to help change, and what you want from the companies developing AI, including Anthropic. There are no right answers; we want to hear about your experiences in your own words… And we don’t want to give too much away before you take it! No. It’s optional. We will publish your complete interview and the country associated with your conversation. We will not include your Claude account information, such as name or email address. Your perspective will be heard not just by us, but by anyone interested in studying and improving AI. We want anyone to be able to study how AI is impacting people and how AI development should progress. Researchers and policymakers around the world can read interviews in their own languages, read responses shared from different regions, and ask questions we didn’t think to ask. Many bring expertise we don't have on staff, and they may have insights we didn’t think to look for. Anyone can check whether our published findings actually reflect what participants said. Your interview also becomes part of the public record of this moment in AI, open to future historians, journalists, and researchers, and can inform how society responds to this moment. That record is also hard for AI companies and others shaping AI development to ignore. If many participants say companies like Anthropic should do something differently, that language stays public where anyone can reference it. While we won’t publish your Claude account information, such as your name or email address, we encourage participants to consider that anyone can read a public interview. This includes employers, friends and relatives, as well as governments and other companies. Depending on what a participant shares, others may be able to link the participant back to their interview. Details that seem harmless on their own, such as age, job, city, or where you studied, can make it easier to identify you. We don’t ask for these details in the interview and encourage participants to avoid sharing them. Once public, an interview could be used in ways a participant does not intend, including by scammers and other people who may seek to misuse personal information. We can remove our copy of your interview on request, but we can’t delete what people have already saved. For that reason, treat your decision to make your interview public as permanent. We won’t attach your name or account information, but we don’t edit interviews, so we can’t guarantee you won’t be re-identified from what you say. Someone could use AI to combine small details, like employers or past projects, to work out who a participant is. Researchers [have shown](https://arxiv.org/abs/2602.16800) this can work with interviews from Anthropic Interviewer. The best protection is to not share information that could identify you or others. During the interview, we show examples of information you shouldn’t share, and you review your full interview before deciding whether to make it public. Before the interview, we explain what making an interview public involves. We tell participants that we will not include their Claude account information in the public release, but that others may be able to identify them from what they share. We also explain that making an interview public means anyone can read, save, and use it. We can remove our copy on request, but we can’t delete what people have already saved. During the interview, you will see examples of the types of sensitive information we advise not to share. After the interview, you can reread what you said before deciding whether to share it publicly. You will be reminded of both the benefits and risks, and if you choose to share, we will confirm your decision one last time. We’ll still analyze it as part of our Societal Impacts research, but your full interview won’t be published. We may still publish and share aggregated or de-identified findings resulting from your interview. If you don’t want us to analyze it either, you can end the interview early (see “Can I decide to end the interview early and delete the record?” below for more detail). Not fully. Once your interview is public, anyone can read, save, and use it. We can remove our copy of your interview on request, but we can’t delete what people have already saved. Treat the decision as permanent. [Learn more](https://privacy.claude.com/en/collections/19084219-anthropic-interviewer). No, we don’t plan to edit interviews before publishing them. Once the study is complete, we’ll do a final review and may remove interviews we don't believe would be responsible to publish, such as interviews with confidential information or ones that violate our [Usage Policy](https://www.anthropic.com/legal/aup) or [Terms of Service](https://www.anthropic.com/legal/consumer-terms). We'll analyze interviews as part of our Societal Impacts research, share our findings publicly, and use the results to inform how we develop our models and products. [Learn more](https://privacy.claude.com/en/collections/19084219-anthropic-interviewer). Yes, you can stop at any time by clicking the “End interview” button. If you end early, your partial interview will be permanently deleted, and won’t be analyzed or published, including by Anthropic. This study will run from September 29 to October 6, 2026, and is open to Free, Pro, and Max users on Claude and Claude Code whose accounts are at least two weeks old. Each interview takes roughly 15 minutes and is conducted by [Anthropic Interviewer](https://www.anthropic.com/research/anthropic-interviewer), an AI that asks you questions. We updated the questions, the study is open on more Claude products, and for the first time you can choose to make your interview public. The interviews will be published online once we have completed the study and our initial analysis. Participants in this study will all be people who use Claude, which is not a representative sample of the public. Published interviews are a further subset, since people who opt in for public release may differ from those who don’t. And as with any interview, the questions we ask shape what people share. Even so, nothing like this has been public before. We believe it can significantly advance the study of how AI is changing people’s lives.