Artificial intelligence has significantly transformed the nature of work and altered the way people engage in it. University students who are about to enter the work-force face a future in which artificial intelligence may replace their jobs and threaten their careers. To mitigate this risk, it is crucial for universities to encourage students to study artificial intelligence and equip them with the necessary skills. Enrolling in artificial intelligence courses is an achievement-related decision, and future-oriented emotions can strongly influence students’ choices. Although expectancy-value theory is widely used to explain academic motivation and behavior, it overlooks the influence of anticipated emotions. To fill this void, this study extends expectancy-value theory by incorporating anticipated regret to investigate the antecedents of artificial intelligence learning intention. Data were collected from 191 university students via an online questionnaire and analyzed using multiple regression analysis. The results reveal that self-efficacy positively influences intention to learn artificial intelligence. Intrinsic value and utility value have positive effects on intention to learn artificial intelligence, whereas effort cost hinders the intention. Anticipated regret plays a key role in shaping intention to learn artificial intelligence. This study provides valuable insights for universities and instructors and enriches literature on artificial intelligence learning intention.