Anthropic Public Record Survey Reveals Widespread Bipartisan Fear of AI Job Loss and Demand for Government Regulation

Anthropic on Friday released the findings from its first "Anthropic Public Record" survey, offering a detailed look at how Americans view artificial intelligence. The nationally representative poll, conducted in November and December of 2025 with 51,993 respondents, finds that Americans are largely united in their hopes and anxieties about AI. Across party lines, there is strong backing for government oversight, and remarkably little faith in the companies building the technology.

Top Hopes: Curing Disease and Helping the Disabled

When asked to rank their top three hopes for AI from a list of 17 potential benefits, nearly half of Americans, 48%, picked curing diseases like cancer or Alzheimer's. Helping people with disabilities came second, chosen by 36% of respondents. Two hopes tied at 23%: accelerating technological progress and making everyday life easier.

Lower-ranked hopes included therapy, reducing loneliness, and other applications where AI might substitute for human contact. The pattern suggests the public sees AI as a tool to augment human capability, not replace human interaction.

Top Fears: Job Loss, Cognitive Dependency, and Misinformation

The survey measured concern across 20 possible AI harms. AI-induced job loss was the most widespread fear, cited by 64% of Americans, and it was the number one fear in every state surveyed. The second most common fear was cognitive dependency (56%), the worry that reliance on AI will erode people's ability to think independently. Misinformation followed at 52%.

Anthropic noted that the most common fears, job loss, cognitive dependency, misinformation, criminal use, and surveillance, tend to be near-term and concrete. Each has a precedent in earlier technologies like automation, smartphones, and social media. Americans, the survey found, were generally more worried about deliberate misuse of AI than about the technology "going rogue."

Job loss concerns were notably higher among Americans with more education, nearly 10 percentage points higher among those with postgraduate degrees compared to those with a high school education or less. But people who reported using AI at work every day were significantly less worried (54%) than those who never used AI (70%), suggesting that hands-on experience with AI may reduce perceptions of threat.

Strong and Bipartisan Support for Government Regulation

Over 70% of Americans believe the government should play a role in regulating AI. The support is broadly bipartisan: 79% of Democrats, 68% of Republicans, and 69% of Independents. A majority in every state and territory backed government involvement, from a high of 81% in Washington, D.C., to a low of 63% in Hawaii.

When asked about specific domains for government action, privacy (56%), child safety (52%), and liability for harm (49%) drew the highest support.

Public Sees Liability and Safety as Key Levers

To ensure AI benefits humanity, respondents ranked two actions highest: holding AI companies legally liable for harm (47% chose this among their top three) and prioritizing safety over growth (44%). Independent watchdogs with real power (29%) and slowing development for safety (27%) followed.

Trust in AI companies was the lowest across all institutions tested: only 15% of Americans said they trust AI companies to make decisions about AI development and use. That figure trailed the federal government (20%), state and local government (19%), and independent experts (43%).

Heavy AI Users: A Glimpse of the Future

The survey identified a group of "integrated users", about 6% of Americans who use AI daily for both work and personal life. These users skew young, male, urban, employed, and college-educated. They are less worried about AI harms than the general public, though Anthropic notes this likely reflects the generally more optimistic outlook of early adopters.

Notably, integrated users support government involvement in AI at essentially the same rate as the national average (74% versus 71%), and their preferences across eight specific governance domains were nearly indistinguishable from the broader public's.

Anthropic plans to repeat the Anthropic Public Record survey regularly, tracking how public attitudes evolve as AI capabilities advance and adoption deepens. The company also intends to expand the survey internationally in future waves.