Skip to content
Longterm Wiki
Back

Humans Are More Likely to Believe Messages from AI (Stanford HAI)

web

Credibility Rating

4/5
High(4)

High quality. Established institution or organization with editorial oversight and accountability.

Rating inherited from publication venue: Stanford HAI

Relevant to AI safety discussions around misuse, disinformation, and the social risks of deploying persuasive AI systems; underscores why transparency and content labeling policies matter.

Metadata

Importance: 55/100news articlenews

Summary

A Stanford HAI study examines how people respond to messages they believe are generated by AI versus humans, finding that individuals tend to place higher credibility or trust in AI-generated content. This has significant implications for misinformation, persuasion, and the societal risks of AI-generated communication at scale.

Key Points

  • People are more likely to believe or trust messages when they are told the source is AI rather than human.
  • This credibility bias could be exploited for disinformation campaigns or manipulation at scale.
  • The finding raises concerns about AI's potential to amplify persuasion and undermine critical evaluation of information.
  • Results suggest current public mental models of AI may inadvertently confer unwarranted authority to AI outputs.
  • Has policy implications for AI-generated content labeling, disclosure requirements, and media literacy.

Cited by 1 page

PageTypeQuality
AI Capability Threshold ModelAnalysis72.0

Cached Content Preview

HTTP 200Fetched Apr 9, 20260 KB
404
Resource ID: 9fc081c471fb3bb0 | Stable ID: sid_PdkpIq8GTd