Anthropic has initiated a new series of surveys named Anthropic Public Record to gather insights on public opinions and feelings regarding artificial intelligence (AI). This first round of surveys, conducted in November and December 2025, engaged close to 52,000 Americans and revealed some intriguing findings.
The survey highlighted that almost half of the respondents (48%) considered finding cures for diseases such as cancer and Alzheimer’s among their top three aspirations for AI. This was followed by aiding individuals with disabilities (36%) and enhancing technology for overall convenience (23%).
Concerns about job losses due to AI were the predominant worry, as 64% of Americans expressed this fear across various states. Other significant apprehensions included dependence on AI (56%) and the spread of misinformation (52%).
A considerable majority, over 70%, of those surveyed advocated for government intervention in AI regulation, showing a notable bipartisan agreement. The public specifically pointed to areas like privacy (56%), child safety (52%), and establishing accountability for AI-related harms (49%) as critical domains for government action.
When prompted about the most effective actions to ensure AI benefits society, respondents prioritized legal accountability for AI companies (47%) and the need for safety to take precedence over rapid growth (44%). Trust in AI firms was markedly low, with only 15% expressing confidence in these companies’ decisions regarding AI development and usage.
Interestingly, Americans exhibited uniform views on several AI-related topics, indicating a collective desire to harness the advantages of AI while also fearing potential disruptions. Disagreement among respondents mainly revolved around the degree of their concerns rather than the issues themselves.
This survey complements Anthropic’s ongoing efforts to explore user experiences with Claude, their AI tool. A qualitative study involving 81,000 Claude users was conducted using Anthropic Interviewer to gain comprehensive insights about how people utilize AI. The Anthropic Public Record survey marks the first attempt to capture general public opinions, reaching non-AI users and thereby enabling a more nuanced understanding of attitudes across various demographics.
Anthropic intends to repeat the Public Record survey regularly, evolving the topics covered as societal concerns change, thus tracking shifts in public sentiment toward AI development and its adoption. Plans for extending this research beyond the United States are also in the pipeline.
Regarding the survey methodology, the online survey was representative of the American population aged 16 and older, achieving a sample of 51,993 respondents, with state-level margins of error varying from ±2.6 to ±9.1 percentage points.
Key hopes and fears regarding AI were explicitly defined in the survey. Curing diseases emerged as the leading aspiration, with a solid 48% of participants ranking it highly, while aiding those with disabilities followed at 36%. In contrast, options like therapy and alleviating loneliness were among the least favored.
When it came to fears, job displacement was the top concern at 64%, followed closely by fears of cognitive dependency at 56%, and misinformation issues at 52%. The concerns listed resonated strongly, as they echoed anxieties linked to previous technological advancements, such as automation and smartphone reliance.
The survey also uncovered significant trends regarding job loss fears across various demographics, revealing that this concern uniformly affects different political affiliation groups, households, and states. Interestingly, higher education levels correlated with increased anxiety about potential job displacements. Conversely, among daily AI users, only 54% reported worries about job loss, compared to 70% for non-users.
The study revealed that perceptions of AI capabilities influence acceptance in workplaces. Overall, while many respondents rated AI’s performance positively, a significant portion preferred minimal AI involvement in their tasks.
Cognitive dependency, the second greatest fear, appeared to be more theoretical than real for most. Only a fraction of those worried about it indicated they would suffer significant disruption if AI were suddenly unavailable.
Integrated users of AI—those incorporating it into both work and daily life—comprise about 6% of the population and show less concern across various issues. This demographic tends to include younger, urban, and college-educated individuals who are more open to new technologies.
The survey emphasizes a strong demand for government action in AI policy, with 71% supporting regulatory involvement. Bipartisan support spans all regions, highlighting a collective desire for responsible governance in AI development.
In wanting AI to serve human interests, Americans converged on two critical demands: ensuring AI companies are legally accountable for damages (47%) and prioritizing safety over rapid development (44%). Trust in AI firms was notably low, but integrated users showed marginally greater confidence while still supporting the call for oversight.
Through initiatives like Anthropic Public Record, Anthropic is committed to understanding and addressing public sentiment towards AI. Their current policy frameworks aim to enhance safety and economic preparedness in light of AI advancements, highlighting the need for inclusive dialogue in shaping the future of AI.
