Introduction
Artificial intelligence is now an integral part of daily work life, but how is it truly perceived by those who use it? To answer this question, Anthropic Interviewer was launched, an innovative research tool that conducted in-depth interviews with 1,250 professionals. This large-scale study aims to understand not just how AI is used, but how it is reshaping professional expectations and identities.
Context: Why Anthropic Interviewer?
Millions of people interact daily with models like Claude, but simple chat logs tell only part of the story. The "why" and the emotional "how" behind each prompt are often missing. To bridge this gap, Anthropic developed a system capable of conducting qualitative interviews at scale, maintaining user privacy while gathering direct feedback.
The system operates in three distinct phases:
- Planning: Creates an interview rubric based on specific research goals.
- Interviewing: Conducts adaptive, real-time conversations (10-15 minutes) directly within the chat interface.
- Analysis: Synthesizes results by identifying emerging themes and prevalent sentiments.
General Workforce: Optimism and Adaptation
From interviews with 1,000 workers across various sectors, a predominantly positive picture emerges. 86% of participants state that AI saves time, and 65% report satisfaction with its role.
However, there is a clear distinction between tasks workers want to delegate and those they wish to keep. There is a trend to cede administrative and repetitive work to AI ("automation"), preserving human interactions and high-value tasks for themselves. Interestingly, 48% of respondents are considering steering their careers toward managing and overseeing AI systems rather than direct technical execution.
"It's a tool to me like a computer was, or a type writer was in the day—computers didn't get rid of mathematicians, they just made them able to do more and that is where I see AI going in the best possible future."
Office Assistant, Study Participant
The Challenge for Creatives: Productivity vs. Identity
The group of 125 creative professionals (writers, visual artists, designers) showed complex dynamics. Although 97% acknowledge time savings and 68% an increase in work quality, anxiety about the future is palpable.
Social Stigma and the Illusion of Control
Many creatives use AI "secretly" to avoid peer judgment or brand devaluation. A recurring theme is the struggle to maintain authorial control: while everyone wishes to remain in charge, many admit that AI often drives creative decisions more than they would like to admit.
"I fully understand that my gain is another creative’s loss. That product photographer that I used to have to pay $2,000 per day is now not getting my business."
Creative Director, Study Participant
The Scientists' Dilemma: Utility Without Full Trust
For the 125 scientists interviewed, Anthropic Interviewer revealed a pragmatic yet cautious approach. AI is widely used for writing manuscripts, literature reviews, or coding, but rarely for the "core" of research like hypothesis generation or experimental design.
The main reason is trust: 79% of interviews highlighted concerns about reliability and model "hallucinations." Scientists desire a partner that can generate new scientific ideas, but current tools do not yet seem up to this critical task.
Conclusion
Data collected by Anthropic Interviewer shows a workforce actively negotiating with technology. There is no outright rejection, but an attempt at integration that preserves human identity. This tool marks a step forward in sociological research on AI, allowing user voices to be heard at an unprecedented scale. For technical and methodological details, you can consult the official Anthropic post.
FAQ
What exactly is Anthropic Interviewer?
Anthropic Interviewer is a research tool developed to conduct automated qualitative interviews at scale, allowing for a better understanding of how professionals interact with AI and what their future perspectives are.
What are the main concerns of creatives revealed by the study?
Despite increased productivity, creatives fear the social stigma associated with AI use, the economic devaluation of their work, and the loss of control over the artistic decision-making process.
Do scientists trust AI for research?
Currently, scientists use AI primarily for support tasks like coding or writing. Trust in using AI for hypothesis generation or core experiments remains low due to the risk of hallucinations.
How can I participate in Anthropic Interviewer studies?
Some Claude.ai users (Free, Pro, or Max plans signed up for at least two weeks) might see a pop-up inviting them to participate in the study. Participation is voluntary and helps shape future model development.
What is the difference between automation and augmentation found in the study?
In interviews, 65% of professionals describe AI use as "augmentation" (collaboration), while only 35% as pure automation. However, analysis of actual chat logs shows a more even split (47% augmentation, 49% automation), suggesting a discrepancy between perception and actual use.