Study Evaluates Trust in AI-Generated Health Advice (TAIGHA) Scale

PLOS Digital Health published a development and validation study describing the Trust in AI-Generated Health Advice (TAIGHA) scale and its 4-item short version, TAIGHA-S, as instruments for measuring users’ trust and distrust in health advice generated by artificial intelligence.

The authors wrote that artificial intelligence (AI) tools, including large language models (LLMs), are increasingly used by the public for health information and health-related decisions. Because users may follow or reject AI-generated advice, the authors said trust in this setting has potential clinical, safety, and healthcare system implications.

The study used a generative AI approach to create use-case-specific candidate items, followed by automated item reduction and a 3-step manual validation process. Content validation included 10 domain experts, face validation included 30 lay participants, and psychometric validation included 385 UK participants who received AI-generated health advice in a symptom-assessment scenario. Participants first selected whether they would seek emergency care, nonemergency care, or self-care, then received AI-generated advice and made a second selection.

After automated item reduction, 28 items were retained and then reduced to 10 based on expert ratings. The final TAIGHA scale showed strong content and face validity, with a scale-level content validity index, average (S-CVI/Ave) of 0.99, and scale-level face validity index, average (S-FVI/Ave) of 0.99. Confirmatory factor analysis supported a 2-factor model measuring trust and distrust, with comparative fit index (CFI)=0.98, Tucker-Lewis index (TLI)=0.98, and standardized root mean square residual (SRMR)=0.03. Internal consistency was high for both trust and distrust subscales.

The TAIGHA-S correlated highly with the full scale (r=0.96) and also showed high reliability. The authors concluded that TAIGHA and TAIGHA-S may be used to assess state trust and distrust in AI-generated health advice, while noting that future studies should validate the tools across languages, cultures, populations, and real-world clinical contexts.

Sources: Kopka M, Majeed A, Spinelli G, El-Osta A, Feufel M. The trust in AI-generated health advice (TAIGHA) scale and short version (TAIGHA-S): development and validation study. PLOS Digit Health. 2026;5(7). doi:10.1371/journal.pdig.0001488