Analysis of Anti-AI Sentiment and the Uncanny Valley Effect

Survey Data on Public vs Expert Views

Pew's 2025 survey found 76% of AI experts said AI would benefit them personally, while only 24% of the U.S. public said the same. The public was much more likely to say AI would harm them than benefit them.

Negative sentiment appears to be growing. Quinnipiac found in March 2026 that 55% of Americans thought AI would do more harm than good in their day-to-day lives, up from 44% in April 2025. The same survey found 64% thought AI would do more harm than good in education.

Reasons for Public Hostility

• Fraud
• Misinformation
• Privacy invasion
• Concentration of power
• Job displacement (which carries emotional weight by threatening status, livelihood, and social usefulness)

The Uncanny Valley Framework

Masahiro Mori introduced the uncanny valley concept in 1970, originally describing how corpses and zombies sat deep in the valley. The original graph tied human likeness directly to revulsion once the likeness crossed into something animate-seeming and lifeless.

Current literature offers several explanations for why this drop in affinity happens, including:

• Mismatch
• Category ambiguity
• Expectation violation
• Disgust
• Threat-related mechanisms

How AI Triggers Uncanny Reactions

AI now appears in forms that trigger human expectations throughout daily life:

• Text that sounds conversational
• Voices that sound natural
• Images and video that almost pass
• Agents that mimic competence, memory, initiative, or empathy

Mismatch is the strongest basic explanation. AI presents cues that invite human social expectations, then fails to satisfy them reliably:

• Natural language invites expectations of understanding
• Warm tone invites expectations of care
• Realistic video invites expectations of authenticity
• Agentic behavior invites expectations of judgment and competence

Research on Repeated Exposure

Some work on repeated interaction with robots suggests uncanniness can decrease with familiarity in certain contexts. Familiarity can reduce startle while leaving behind a more stable sense that the category is untrustworthy. This fits AI especially well because people encounter many versions of the same near-human pattern across modalities.

Disgust and Danger Mechanisms

A 2025 study on virtual agents explicitly frames findings in terms of the pathogen-avoidance hypothesis. Moosa and Ud-Dean argue that pathogen avoidance alone is too narrow because even a fresh corpse can provoke strong aversion before visible decay appears. Their proposal is that the uncanny valley reflects a danger-avoidance system more generally.

AI often presents near-human abnormality that could fit this account: it speaks with confidence without understanding, performs social fluency without satisfying conditions that make human social signals trustworthy. This mismatch could recruit disgust or danger-detection processes even when the stimulus is text, voice, or video rather than a literal body.