The abstract is read at the publisher; the summary is JClub's.
At the highest difficulty in an fMRI study of 44 participants, trusting an autonomous system was linked to a 51% trust probability versus 72% for a human partner despite equal 52% accuracy.
- Why it matters: It was unclear whether active trust decisions for AI systems use the same brain evaluative machinery as trust in people, and whether neural engagement differs when reliability is uncertain.
- What they did: Participants repeatedly chose to trust or distrust a human or autonomous partner with matched prescripted accuracy across five difficulty levels while undergoing fMRI, enabling comparisons of trust-related neural patterns and coupling.
- The result: Distrust recruited a prespecified performance monitoring network for both partners, but trust-related engagement was attenuated for the autonomous system; trust in the system tracked reliability without exceeding it.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.