Anthropic Interviewer is a new Claude-based research tool that automates large-scale, semi-structured interviews to understand how people use and feel about AI at work. After reading the study, I think it points to an AI conundrum at work. People are using AI, but they do not feel safe saying so.
Anthropic Interviewer is “designed to run deep, qualitative conversations at scale and then feed results back to human researchers for analysis.”
Anthropic says Interviewer runs in three stages:
- Planning: it creates an interview rubric and question flow, guided by a system prompt and reviewed by human researchers.
- Interviewing: it conducts 10-15 minute, adaptive interviews inside Claude.ai.
- Analysis: it helps a human researcher synthesise transcripts, and they also use a separate automated tool to cluster themes and quantify how often topics appear.
The interviews are designed to feel more like conversations than surveys, with the AI adjusting follow-up questions based on each participant’s responses while still following a consistent research protocol.The initial 1,250-interview study
For the first test, Anthropic used the tool to conduct 1,250 interviews across three cohorts:
- 1,000 general workforce participants
- 125 creatives
- 125 scientists.
Each interview lasted around 10-15 minutes and explored how participants use AI in their day-to-day work, what they find most useful, and what worries them about the technology. Anthropic says they are publicly releasing the interview data with participant consent.
Five key findings
1. AI boosts productivity — but feelings are mixed
Most people see AI as a productivity gain, but they remain conflicted. In the general workforce sample, 86% said AI saves them time, and 65% said they are satisfied with the role AI plays at work.
2. “AI creeping” is real: stigma is shaping behaviour
Anthropic reports that 69% of professionals mentioned social stigma around using AI tools at work. In short, people are using AI, benefiting from it, and still feel they need to hide it.
3. Identity protection is a core theme
Workers want to offload routine tasks while keeping the parts of their roles that define them professionally. Many imagine a future where they oversee AI, rather than do everything themselves.
4. Creatives: efficiency now, anxiety later
Creatives report using AI to speed up their work, while also grappling with peer judgement, fears of displacement, and questions about what happens to “human” creative identity.
5. Scientists: they want an AI partner, but trust is the blocker
Scientists consistently want help with hypothesis generation and experiment design, but describe current use as more peripheral (writing, debugging, summarising) because they do not trust models enough for core research work.
Want more support with AI in your organisation, or to boost your career? Get in touch.
Book a call in my calendar.
Get a copy of AI Playbook.
About the Podcast Host
Joanne Sweeney is a specialist in public-sector marketing and digital communications. She’s the author of Public Sector Marketing Pro and the creator of the Professional Diploma in AI for Public Sector Marketing. With years of experience in training, consulting, and advocating for digital transformation, Joanne provides insight and guidance for government agencies navigating the digital landscape. She is also the host of the award-winning Public Sector Marketing Show podcast.
