AI Trust Scale

What it measures

Measures consumer trust in AI-based systems and intelligent agents across three dimensions derived from Mayer et al.'s (1995) trust model: ability (the AI can do what it promises), benevolence (the AI acts in the user's interests), and integrity (the AI follows principles the user endorses). Increasingly used in research on AI assistants, recommender systems, and autonomous agents.

Response format

Likert (1–7, strongly disagree to strongly agree)

Scale items

Cognitive Trust in AI — reliability and competence beliefs (Likert 1–7, strongly disagree to strongly agree)

• I believe [AI system] is reliable.
• I believe [AI system] is competent.
• [AI system] performs consistently.
• [AI system] gives accurate information.
• I can count on [AI system] to do what it says it will do.

Affective Trust in AI — feelings of security and comfort

• I feel comfortable using [AI system].
• [AI system] makes me feel at ease.
• I feel secure interacting with [AI system].
• I feel [AI system] has good intentions toward me.
• I feel [AI system] cares about my needs.

Full scale items shown. Always cite the original source when using this scale in research.

Usage notes & scoring

Three subscale scores and total trust score. Score = mean per subscale. Increasingly relevant as AI-powered recommendations, chatbots, and autonomous agents proliferate in consumer contexts. Related measures include algorithmic aversion (Dietvorst et al., 2015) and AI acceptance (adapted TAM).

Recommended companion scales

Citation

Söllner, M., Hoffmann, A., & Leimeister, J.M. (2016). Why different trust relationships matter for information systems users. European Journal of Information Systems, 25(3), 274–287.

Keywords