Answers are grounded in this paper and cite their sources.
As artificial intelligence (AI) systems become more autonomous and trustworthy, users may invest less effort in monitoring systems and shift responsibility away from themselves. This study examines how system reliability as an indicator of trustworthiness and level of automation (LOA) influence responsibility attributions before and blame after a negative outcome. In a 2 (reliability: lower vs. higher) × 2 (LOA: lower vs. higher) between-participants design (N = 261), participants completed a personnel selection task where they selected applicants for interviews with the assistance of an AI system. The results indicate that higher LOA was associated with lower internal and higher external responsibility and blame attributions, with patterns consistent with mediation via perceived control and information processing. Higher system reliability showed a similar pattern for internal responsibility, consistent with mediation via information processing. These findings increase our understanding of a gap in current research: how high LOA and high system reliability can undermine users’ perceived control and information processing potentially through changes in monitoring efforts, leading them to offload responsibility and blame. To counteract this, AI systems should not only provide transparency and control but also implement strategies that actively maintain user engagement in the decision-making process.
No takes yet. Share an insight, caveat, or question.
Kares et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: