Measuring trust in artificial intelligence: validation of an established scale and its short form.

An understanding of the nature and function of human trust in artificial intelligence (AI) is fundamental to the safe and effective integration of these technologies into organizational settings. The Trust in Automation Scale is a commonly used self-report measure of trust in automated systems; however, it has not yet been subjected to comprehensive psychometric validation. Across two studies, we tested the capacity of the scale to effectively measure trust across a range of AI applications. Results indicate that the Trust in Automation Scale is a valid and reliable measure of human trust in AI; however, with 12 items, it is often impractical for contexts requiring frequent and minimally disruptive measurements. To address this limitation, we developed and validated a three-item version of the TIAS, the Short Trust in Automation Scale (S-TIAS). In two further studies, we tested the sensitivity of the S-TIAS to manipulations of the trustworthiness of an AI system, as well as the convergent validity of the scale and its capacity to predict intentions to rely on AI-generated recommendations. In both studies, the S-TIAS also demonstrated convergent validity and significantly predicted intentions to rely on the AI system in patterns similar to the TIAS. This suggests that the S-TIAS is a practical and valid alternative for measuring trust in automation and AI for the purposes of identifying antecedent factors of trust and predicting trust outcomes.