A landmark survey of nearly 52,000 Americans conducted by Anthropic reveals a striking gap between public hopes and fears about artificial intelligence. While disease curing ranks as the top hope for AI technology, job loss emerges as the dominant concern across all demographic groups, with 64% of respondents worried about displacement.

Anthropic's inaugural "Public Record" survey, fielded in November and December of 2025, marks the company's first effort to understand how the general public,not just AI users,thinks about the technology. The findings paint a nuanced picture of American attitudes toward AI development and regulation.

What Are Americans' Greatest Hopes for AI?

When asked to identify their top three hopes for AI from a list of 17 possibilities, Americans prioritized tangible health benefits above all else. Curing diseases like cancer and Alzheimer's topped the list, with 48% of respondents ranking it among their top three hopes. This was followed by helping people with disabilities at 36%, and then technological progress and making life easier in general, each cited by 23% of respondents.

Notably, aspirational hopes involving AI as a substitute for human connection, such as therapy or reducing loneliness, ranked lowest among the options presented. This suggests Americans view AI's greatest promise in augmenting human capability rather than replacing human relationships.

Why Do Americans Fear Job Loss More Than Any Other AI Risk?

Job displacement emerged as the most widespread concern in the survey, held by nearly two-thirds of Americans. What makes this finding particularly striking is its consistency across traditional demographic divides. The fear of job loss was the top-ranked concern among both Democrats (67%) and Republicans (62%), in households with and without children, and in every state surveyed, ranging from Iowa at 71% to Mississippi at 57%.

The second most common fear was cognitive dependency on AI, where 56% of respondents expressed concern that AI integration could leave people unable to think for themselves. Misinformation ranked third at 52%. Interestingly, these near-term, concrete harms dominated American concerns far more than abstract risks like AI "going rogue" or misalignment.

However, hands-on experience appears to shift perspectives. People who use AI at work every day are notably less worried about job loss than those who don't use AI at all, with 54% versus 70% expressing concern. This suggests that direct familiarity with AI's actual capabilities and limitations may help workers develop skills to augment rather than be replaced by the technology.

How to Understand Public Sentiment on AI Regulation and Accountability

• Government Oversight Support: Over 70% of Americans surveyed believe the government should play a role in regulating AI, with this support cutting across partisan lines. Americans were most eager for government action in privacy (56%), child safety (52%), and liability for harm (49%).
• Corporate Accountability Demands: When asked what would best ensure AI benefits humanity, Americans ranked holding AI companies legally liable for harm (47%) and prioritizing safety over growth (44%) as the highest-leverage actions.
• Trust Deficit: Only 15% of Americans said they trust AI companies to make decisions about how AI is developed and used, highlighting a significant credibility gap between the industry and the public.

The survey reveals broad consensus on AI regulation across demographic lines. On most questions, responses did not heavily divide Americans along typical partisan, geographic, or educational lines. Instead, disagreement tended to manifest in the intensity of people's views rather than fundamental disagreement about whether AI poses risks or offers benefits.

Anthropic plans to repeat the Public Record survey regularly, evolving its scope as new topics become salient and allowing researchers to track how public attitudes shift as AI model capabilities advance and adoption deepens. The company has indicated plans to expand the survey outside the United States in the future.

This research builds on other ongoing work at Anthropic to understand AI perceptions. The company recently conducted a global qualitative study of 81,000 Claude users through its Anthropic Interviewer tool, which enables in-depth interviews at scale. Anthropic also regularly releases data from the Anthropic Economic Index, which draws on anonymized Claude usage data to show how people around the world are employing AI.

The survey methodology employed rigorous standards to ensure representativeness. Anthropic conducted a nationally representative online survey of 51,993 Americans, sourced from YouGov and weighted to US Census benchmarks. State samples ranged from 232 respondents in Alaska to 1,902 in New York, with state-level margins of error between plus or minus 2.6 and plus or minus 9.1 percentage points.