Prioritize trust calibration over trust maximization when evaluating explainability and other design factors¶

Strategy · All strategies
Evidence · 1 claim (1 mixed) · 1 study (1 associational), q3 · 0 of 1 report an effect size · 1 claim rests on one study

Description¶

The review recommends that research on AI design factors focus on fostering appropriate levels of trust rather than simply increasing trust. It argues that because AI systems still produce errors and hallucinations, "studyingtrustcalibration–ensuringusersrelyonAIappropriately–iscurrentlythemostpracticalandvaluablefocus". Concretely, it suggests evaluating explanations by whether they help users distinguish reliable from unreliable outputs, and comparatively investigating explanation format, timing, detail, and accuracy to clarify when explainability supports appropriate reliance versus overtrust or unnecessary distrust.

Design Implications¶

Context¶

Requirements¶

• Studies should assess whether design features help users distinguish between reliable and unreliable system outputs rather than only measuring trust increases.

Constraints¶

• The review notes this calibration focus is currently not prioritized in the literature, which emphasizes increased trust.

Target Learners¶

• users of AI-enabled systems, including learners using intelligent tutoring systems and conversational AI

Target Learning Goals¶

• appropriate and calibrated reliance on AI-enabled systems

Claims¶

• The trust–appropriate reliance relationship is non-linear, with the decline attenuating at higher trust and no reversal within the observed data [~W]

Related Strategies¶

• Target calibrated trust and anxiety alongside AI literacy instruction when supporting existing generative AI users

Examples¶

-

Key Sources¶

• Abramson SL, Sora-Cardenas J, Nandakumar P, Bruce JE, Gopinath N, Zangari LG and Feijóo-García PG (2026). Trust and interaction design in AI-enabled systems: a systematic literature review. Front. Psychol. 17:1932655. https://doi.org/10.3389/fpsyg.2026.1932655