Today for AI
HOT RADAR
agentHEAT 11.3°

AI CURATED HOT STORY · 10/30/2026

Anthropic Study: Experts Adopt Random AI Opinions as Their Own

1 reports archived1 independent sourcesupdated 10/30/2026, 03:33:37
Synthesis & Latest Updates
1 Sources Cross-Validated

An Anthropic experiment revealed that 50 domain experts adopted random AI agent preferences (determined by coin toss) in up to 86% of analytical tasks, mistakenly believing these were their own independent judgments. This behavior persisted regardless of expertise or trust levels, raising concerns about false consensus weakening AI safety checks despite efficiency gains.

LATEST/An Anthropic experiment revealed that 50 domain experts adopted random AI agent preferences (determined by coin toss) in up to 86% of analytical tasks, mistakenly believing these were their own independent judgments. This behavior persisted regardless of expertise or trust levels, raising concerns about false consensus weakening AI safety checks despite efficiency gains.

Hourly heat curve

3 fully observed hours
Now
11.3
Peak
11.904:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first
  1. The DecoderT2·78 pts

    Anthropic Study: Experts Adopt Random AI Opinions as Their Own

    • Experts are highly susceptible to AI agent guidance in ambiguous tasks, with adoption rates reaching 86%, often unaware that opinions were manipulated.
    • Professional experience and trust in AI do not mitigate this 'opinion assimilation' effect, indicating a universal cognitive vulnerability.
    • Widespread use of the same model may create 'false consensus' in research; agents should argue both sides or solicit user views first.