AI dependency increases trust in incorrect health information, study finds

A study involving two experiments with college students and Amazon Mechanical Turk participants found that learned dependency on generative AI for health information increases trust, even when the information is incorrect. Participants who were more dependent on AI were more likely to trust inaccurate outputs. The research concluded that while accuracy boosts trust, overreliance on AI can lead to poor judgment, and simple text highlighting is not enough to mitigate this issue, suggesting a need for better interface designs to encourage critical evaluation. AI

IMPACT Overreliance on AI for health information can lead to trusting incorrect outputs, highlighting the need for better interface designs to promote critical evaluation.

RANK_REASON Academic paper detailing experimental findings on AI trust and dependency. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• AI-generated health information
• Amazon Mechanical Turk
• Arif Ahmed
• Generative AI
• Learned Dependency
• Trust

• paper
• safety

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →