Results from the first Anthropic Public Record

Details

The Anthropic Public Record is a new recurring survey series tracking American public opinion on AI. The first wave, conducted in November and December 2025 through YouGov with 51,993 Americans weighted to US Census benchmarks, reveals patterns that cut across partisan, geographic, and educational lines.\n\nThe top public hopes: 48% ranked curing diseases like cancer or Alzheimer's among their top three AI hopes, followed by helping people with disabilities (36%) and making technological progress or life easier (tied at 23%). Healthcare remains the most resonant public use case for AI benefits.\n\nThe top fears: AI-induced job loss was the dominant concern in every single US state, cited by 64% of respondents. Cognitive dependency (56%) and misinformation (52%) followed. Notably, job fear was distributed broadly — not concentrated in any particular demographic or political group — which gives it unusual political weight.\n\nOn regulation: 70%+ support government involvement in AI, with strongest demand in privacy (56%), child safety (52%), and liability for harm (49%). Only 15% of Americans trust AI companies themselves to make decisions about AI development — a striking trust deficit given that this survey was commissioned by one of those companies.\n\nThe survey will repeat regularly and expand internationally, creating a longitudinal dataset for tracking public attitudes as model capabilities advance.