Understanding how artificial intelligence is reshaping work has largely relied on usage statistics or small-scale surveys. Anthropic is introducing a new method for this research with Anthropic Interviewer, an AI-powered tool designed to conduct and analyze detailed conversations at an unprecedented scale.

To test the system, the company used it to interview 1,250 professionals—including a general workforce sample, scientists, and creatives—about their experiences and feelings toward AI. The findings, now publicly released, reveal a workforce that is largely optimistic but navigating significant anxieties about identity, peer judgment, and the future of their roles.

Key Findings from 1,250 Interviews

The research, which allowed for open-ended conversation rather than multiple-choice questions, uncovered nuanced perspectives across professions:

· General Workforce: A majority (65%) view AI as an augmentative tool that collaborates with them, not just an automation engine. While 86% reported AI saves them time, over half expressed anxiety about its long-term impact. Many envision a future where they manage AI systems, focusing on human oversight and interpersonal tasks while routine work is automated.

· Creative Professionals: Nearly all creatives (97%) said AI boosted their productivity, and 68% reported an increase in work quality. However, 70% acknowledged managing stigma from peers about using AI, and many voiced deep economic anxiety about displacement and market flooding. A universal theme was the desire to maintain ultimate creative control, even as some admitted AI sometimes drives creative decisions.

· Scientists: Researchers reported using AI primarily for auxiliary tasks like writing, coding, and literature reviews, not for core research functions like hypothesis generation or experimental design. The primary barrier was trust and reliability, cited by 79% of scientists, who noted that constantly verifying AI outputs often negates the promised time savings. Despite current limitations, 91% expressed a desire for more capable AI research partners in the future.

How Anthropic Interviewer Works

The tool represents a novel application of AI for social science. It operates in three stages:

1. Planning: Human researchers collaborate with Claude to develop an interview rubric and question set.

2. Interviewing: The AI conducts real-time, adaptive text-based interviews that last 10-15 minutes, following the research plan while allowing for natural conversation tangents.

3. Analysis: The tool helps researchers analyze the transcripts, identifying emergent themes and quantifying their prevalence across the dataset.

Implications and Availability

Anthropic positions this tool as a way to "center human voices" in AI development, creating a feedback loop between user experience and model improvement. The company has begun partnerships with teacher unions, scientific institutions, and creative organizations to conduct targeted research.

The anonymized interview data from this study has been made public for broader academic and societal analysis. A pilot version of Anthropic Interviewer is now accessible to some Claude.ai users, inviting them to participate in ongoing research about AI's role in their lives and work.

This approach highlights a growing trend of using AI not just as a subject of study, but as an active instrument for large-scale qualitative research, potentially offering richer insights into technology's societal impact than quantitative metrics alone.