Americans Fear AI Will Cost Jobs, Hope It Can Cure Diseases

Anthropic conducted its first Public Record survey of 51,993 U.S. adults in late 2025 and found 64% of respondents fear artificial intelligence will lead to job losses. Forty-eight percent placed curing diseases such as cancer or Alzheimer’s among their top three hoped-for applications for AI.

Job-loss concern topped every state, ranging from 71% in Iowa to 57% in Mississippi. The worry was recorded by 67% of Democrats and 62% of Republicans. Respondents with postgraduate degrees were roughly 10 percentage points more likely to report fear of job loss than those with a high school diploma or less. People who use AI at work every day reported lower concern about losing a job — 54% versus 70% for those who do not use AI at all.

Other leading fears included cognitive dependency at 56% and misinformation at 52%. Trust in AI companies was low: 15% of respondents said they trust AI firms to steer the technology. Anthropic wrote that this was the lowest trust score for any institution tested, below the federal government at 20%, state and local government at 19%, international bodies at 20%, and independent experts at 43%.

Data tracked around job cuts show AI was cited in 38,579 U.S. job eliminations in May. For 2026, employers have linked 87,714 cuts to AI so far, a total that exceeds the 54,836 layoffs tied to AI in 2025. In Washington, Senators Elizabeth Warren and Bernie Sanders have urged Congress to pass protections for workers affected by automation.

Jeff Bezos, founder of Amazon and an investor in AI startup Prometheus, which recently raised $12 billion at a $41 billion valuation, said he expects AI to create labor scarcity rather than broad unemployment and suggested household labor participation could shift as a result.

On the hopeful side of the survey, 48% ranked curing major diseases in their top three uses for AI, and 36% placed helping people with disabilities among their top three. Seventy-one percent of respondents supported some form of government oversight of AI, including 79% of Democrats and 68% of Republicans. When asked how to keep AI development aligned with public interest, 47% favored holding companies legally liable for harm and 44% prioritized safety over growth.

Anthropic plans to repeat the Public Record survey as AI adoption increases and described the early results as reflecting public interest in medical and accessibility advances alongside skepticism of the companies building and deploying the technology.