Understanding users' intention to adopt AI avatars in online learning: an integrated model of TAM and DOI

In plain English

A survey of 653 adults with prior digital-learning experience asked which avatar design attributes actually drive willingness to use an AI avatar tutor, testing personalization, feedback, transparency and interactivity through a combined technology-acceptance and innovation-diffusion model. Personalization was the strongest lever, predicting observability, relative advantage, ease of use and usefulness (all p < 0.001), and the model explained 38.9% of adoption intention. Older respondents were less willing to adopt; gender, education and prior avatar experience made no difference. The design is cross-sectional and self-reported, so these are associations with stated intent, not evidence that any feature changes behaviour.