Researchers at the HJ Institute of Technology and Management (HJITM) have published a study analyzing the factors that determine ChatGPT users’ willingness to pay for the paid version (ChatGPT-4) in the International Journal of Human–Computer Interaction (IJHCI, SSCI), an international journal in the field of human–computer interaction.

The researchers analyzed responses from 554 ChatGPT users in Korea, including university students and office workers, using partial least squares structural equation modeling (PLS-SEM). They examined whether perceived intelligence and service quality of ChatGPT lead through usefulness, knowledge management and satisfaction to willingness to pay, and what roles users’ innovativeness and perceived risk play.

The analysis showed that perceived intelligence strongly increased usefulness (β=0.510) and knowledge management (β=0.404), while service quality strongly increased knowledge management (β=0.434) and satisfaction (β=0.376). Usefulness (β=0.237) and knowledge management (β=0.171) raised satisfaction but had no direct effect on willingness to pay. In other words, judging a tool useful is not enough on its own to get users to open their wallets.

What raised willingness to pay was satisfaction (β=0.221) and innovativeness, the tendency to adopt new technologies early (β=0.129). By contrast, perceived risk — the feeling that sharing personal information with ChatGPT is risky — lowered willingness to pay the most (β=−0.261). No moderating effects of innovativeness or perceived risk on the other relationships were found.

The study is significant in showing that the key to converting generative AI users to paid plans lies less in functional usefulness than in overall satisfaction with the user experience and in managing perceived risk. It suggests that AI service companies should raise satisfaction through service quality while clearly communicating their privacy policies to reduce the risk users perceive.

HJITM continues to conduct research on the monetization of generative AI services and user behavior.