SAN FRANCISCO — It isn’t just credit unions and their own workers who are both excited and worried: artificial intelligence users around the world see the technology as both a powerful tool and a source of growing fear, according to a new report by Anthropic based on more than 80,000 interviews conducted using its Claude chatbot.
The company said the study — which gathered responses from about 80,500 users across 159 countries and 70 languages — is the largest qualitative research effort of its kind, capturing detailed, open-ended perspectives on how people use and view AI.
The findings suggest that optimism and concern about AI are not divided into separate groups, but instead often coexist within the same individuals.
“Hope and alarm didn’t divide people into camps,” the report said, with users frequently expressing both enthusiasm for AI’s benefits and cautions over its risks.
Key Themes
According to Anthropic, key themes identified in the report include:
- Productivity and opportunity: Many respondents said AI helps them save time, improve work quality and pursue goals such as financial stability or professional advancement.
- Job displacement concerns: Some users reported losing jobs or fearing replacement as companies adopt AI systems.
- Reliability and control: The most common concern was that AI systems may produce incorrect results or reduce users’ sense of control.
- Dependence and skill erosion: Participants expressed worry that reliance on AI could weaken critical thinking or professional skills.
How Interviews Were Conducted
Anthropic conducted the interviews over a one-week period in December using an automated “Interviewer” version of Claude, which asked users structured questions and follow-ups in conversational form.
The company said the approach allows researchers to collect qualitative data at a scale previously not possible, combining depth of responses with global reach.
The report also found that many users view AI as a “thought partner” rather than a replacement for human work, using it for tasks ranging from contract review and education to personal problem-solving.
Longer-Term Concerns
At the same time, concerns about long-term societal effects remain widespread, including fears about overreliance, loss of agency and broader economic disruption, Anthropic reported.
Anthropic said the findings highlight the need for clearer definitions of what “AI going well” should look like, grounded in real-world user experiences rather than abstract debates.
The study reflects the views of active users of the company’s Claude platform, rather than the general population, and is intended to inform future AI development and policy discussions, the company said.
