We investigated how predictions regarding the outcome of sports competitions are affected by concurrent predictions which were believed to originate from either a human sports expert or an AI system. Participants were asked to estimate the probability of a certain team to win a group match against the respective opponent in various matches of the 2024 Olympic Games and to report their confidence in that estimate. Moreover, they received numerical estimates for that probability which were either high or low, and which were believed to originate from a human expert or an AI system. Participants’ estimates were biased towards these supplementary estimates, thus reflecting an anchoring effect. Yet, this effect’s size was independent of the believed source of the anchor and even non-significantly higher for AI support. Confidence in provided judgments was even significantly higher when the believed origin of supplementary estimates was AI rather than human. Altogether, the results suggest no bias against AI-generated sports outcome predictions, highlighting the task-dependency of reliance on AI.
No takes yet. Share an insight, caveat, or question.
Wunderlich et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: