2025/11/04 by Zdenek Smutny, Zdeněk Smutný, Frantisek Sudzina +1
Computer Science · Medicine · Engineering · #AI in Service Interactions #Artificial Intelligence in Healthcare and Education #Robotic Process Automation Applications
paper · doi:10.1080/10447318.2025.2573037
Generative artificial intelligence (AI) tools are reshaping individual work performance. While many studies explore the adoption of generative AI tools, few examine mediations and factors influencing performance expectancy and intentions to use AI chatbots like ChatGPT. This study builds on Camilleri’s (2024) framework, integrating enhanced UTAUT and IAM models, and presents extended replication. A survey questionnaire (N=787, aged 18–34 y) was analyzed using SmartPLS4. The results show that performance expectancy significantly mediates the relationship between effort expectancy, source trustworthiness, and information quality to intentions to use AI chatbots. Sex differences were found, with men prioritizing the quality of information, while women emphasize the source trustworthiness. Compared to replicated research, distinct cultural and demographic factors influenced adoption outcomes. Despite user-intuitive control similar to internet-mediated human communication, which facilitates adoption among young people, users remain cautious due to risks like hallucinations, social bias, misinformation, or adversarial prompts. The study contributes theoretically and empirically to understanding how AI chatbots affect work-related behavior and decision-making.