Anthropic Public Record: US Public Sentiment on AI

Anthropic has released the results of its first Anthropic Public Record, a large-scale survey of nearly 52,000 Americans conducted in November and December 2025. The findings reveal a public that is broadly hopeful about AI's potential to cure diseases but deeply concerned about economic disruption and corporate accountability, with a bipartisan consensus supporting government regulation.

Public Hopes and Fears for AI

Americans prioritize medical breakthroughs and accessibility as the primary benefits of AI, while fearing concrete, near-term societal harms over theoretical alignment risks.

Primary Hopes: Curing diseases such as cancer or Alzheimer’s was the top hope for 48% of respondents, followed by helping people with disabilities (36%), and technological progress/making life easier (tied at 23%).
Primary Fears: Job loss is the most prevalent concern, cited by 64% of Americans. Other significant fears include cognitive dependency (56%) and misinformation (52%).
Nature of Concerns: The public is more concerned with the misuse of AI (criminal use, surveillance, and terrorism) than with AI "going rogue" or misalignment.

Analysis of AI-Induced Job Loss

Fear of job displacement is a consistent concern across almost all demographic lines, though it varies based on education and usage levels.

Demographic Distribution: Job loss is the top fear for Democrats (67%) and Republicans (62%), and is present in every state surveyed, ranging from 71% in Iowa to 57% in Mississippi.
Education Correlation: Concern increases with education level; those with postgraduate degrees are nearly 10 percentage points more worried than those with a high school education or less.
Usage Correlation: Daily AI users at work are notably less worried about job loss (54%) compared to those who do not use AI at all (70%).

Cognitive Dependency and User Experience

While cognitive dependency is a major fear, the data suggests it is currently an anticipatory concern rather than a lived reality for most.

Anticipatory Nature: Of the 56% of Americans worried about cognitive dependency, only about 1/5 would feel significant disruption if AI became unavailable tomorrow. In contrast, 1/3 of those not worried about dependency would feel significant disruption.
Occupational Trends: Educators are among the most worried about dependency, second only to those in arts and design. This aligns with qualitative research showing educators are 2.5 to 3 times more likely to report witnessing cognitive atrophy in students.
Usage Impact: Worry regarding dependency drops by 16 percentage points for those who use AI daily at work (46%) compared to non-users (62%).

AI Governance and the Trust Deficit

There is a bipartisan supermajority supporting government oversight and a profound lack of trust in AI companies to self-regulate.

Support for Regulation: 71% of Americans believe the government should regulate AI, including 79% of Democrats and 68% of Republicans.
Priority Areas for Government Action: The public is most eager for government intervention in privacy (56%), child safety (52%), and liability for harm (49%).
Industry Accountability: To ensure AI benefits humanity, respondents ranked holding AI companies legally liable for harm (47%) and prioritizing safety over growth (44%) as the highest-leverage actions.
Trust Levels: Only 15% of Americans trust AI companies to make decisions about AI development and use. This is lower than trust in the federal government (20%), state/local government (19%), and international bodies (20%).

Insights from "Integrated Users"

Integrated users—those who use AI daily for both work and personal life (approximately 6% of Americans)—serve as a potential preview of future mainstream adoption.

Profile: These users skew young, male, urban, employed, and college-educated, with two-thirds identifying as early adopters.
Attitudes: While integrated users are generally less worried about AI harms and more trusting of institutions, their appetite for government regulation (74%) is nearly identical to the general public (71%).

Methodology

The survey was a nationally representative online study of 51,993 Americans aged 16 and over, conducted by YouGov between November 1 and December 11, 2025. Data were weighted to US Census benchmarks for state, age, gender, education, and race/ethnicity. National margins of sampling error are ±0.6 percentage points at the 95% confidence level.