Trust in AI: Building a Reliable Tool for Online Learners

A new study has created and tested a tool that measures how much students trust AI in online courses. The researchers first gathered ideas from existing research, then asked experts to check the items for relevance. They ran two rounds of statistical tests: one to explore how the questions group together and another with a separate group to confirm the structure. A total of 837 students answered the survey, split into three samples for each analysis step.The exploratory phase revealed five distinct themes that together explain 63. 20 % of the variation in responses. The confirmatory phase, using an independent group, showed a good fit with standard indicators (CFI = . 95; TLI = . 94; RMSEA = . 06). Reliability was strong, with Cronbach’s alpha and omega both at . 94. The scale works similarly for male and female students, and the researchers found no link between trust levels and grades or how often students used the system. These findings highlight that trust is a separate mindset that can shape how learners interact with AI tools. The 21‑question scale offers educators a dependable way to assess trust and can guide the design of future AI‑enhanced learning experiences.