Researchers at the HJ Institute of Technology and Management (HJITM) have published a study analyzing what keeps people using smartphone AI voice assistants in Universal Access in the Information Society (UAIS, SSCI·SCIE), an international journal in the field of information accessibility.

The researchers analyzed responses from 256 Korean users of AI voice assistants using partial least squares structural equation modeling (PLS-SEM). They examined how interaction (a natural back-and-forth conversation), novelty, the attractiveness of the voice, and discomfort experienced during use affect attitude and continuance intention.

The analysis showed that attitude had the greatest effect on continuance intention (β=0.555). Interaction was the only factor that raised both attitude (β=0.241) and continuance intention (β=0.316). Novelty (β=0.416) and voice attractiveness (β=0.170) improved attitude but had no direct effect on continuance intention.

Discomfort worked in the opposite direction. Discomfort such as the awkwardness of speaking to an AI in public did not affect attitude, but it directly lowered continuance intention (β=−0.096). Even if an assistant makes a good impression, people end up using it less if it is uncomfortable to use.

The study is notable for explaining the continued use of AI voice assistants in a way that reflects the distinctive characteristics of voice as an interface. It suggests that developers should focus on improving the quality of interaction so that conversations flow naturally, and on reducing the discomfort users feel in usage situations, rather than on new features or voices.

HJITM continues its research on the acceptance and continued use of AI technologies.