Human-AI Interaction¶
Study of how humans interpret, trust, and act on AI-generated outputs, particularly in decision-critical contexts. In the misinformation domain, this includes how users calibrate reliance on automated credibility indicators, how they update beliefs when shown AI predictions, and how interface design and framing shape their behavior.
Key phenomena include over-reliance (users trust weak AI too much), under-reliance (users dismiss accurate AI advice), and explanations (how transparency about AI reasoning affects adoption).
In information cascades, human-AI interaction becomes especially important: a single user's judgment of an AI prediction can shape public history and influence dozens of downstream users who never see the raw AI signal.
Key papers¶
- One AI Signal, Many Human Judgments: A Bayesian Cascade Analysis of AI-based Credibility Indicators in Online Information Spread — Bayesian analysis of how shared AI signals reshape user judgments and public consensus
Related topics¶
- AI-based Credibility Assessment (AI systems in misinformation detection)
- Information Cascades (social dynamics of information)
- Content moderation (platform interventions)