Emotions towards algorithmic policy decision-making: direct and compensatory associations with trust in governance

Governance is rapidly digitising, yet how citizens’ emotions towards algorithmic policy decision-making relate to their trust in governance institutions remains understudied. Drawing on the affect-as-information model, we argue that emotions constitute a distinct input into institutional trust judgements, operating both directly and by conditioning the weight of cognitive evaluations. We test this argument with a vignette-based survey of 586 citizens. Positive emotions towards algorithmic decision-making were positively associated, and negative emotions negatively associated, with trust in governance institutions, beyond perceived governance performance and satisfaction with public services. Positive emotions further moderate the outcome-trust relationships in a compensatory pattern, whereas negative emotions showed no moderating role. The findings position citizens’ emotions towards algorithmic governance as an independent correlate of institutional trust, with implications for the accountability of algorithmic governance systems.