AI Insight
This study examined how personality traits and experience with text-based generative AI interact to influence users' trust and dependence on these systems. Using survey data from 502 participants and latent profile analysis, researchers identified three personality types with distinct patterns: Well-Adjusted users showed highest dependence and self-efficacy with high experience, Outgoing-Unstructured users displayed highest trust overall, and Emotionally-Sensitive users exhibited the most cautious engagement across all measures. The findings reveal that personality and experience interact significantly to shape how users rely on and trust generative AI systems.
Why it matters
These results suggest that one-size-fits-all approaches to AI integration may be inadequate, and personalized strategies based on personality profiles could help mitigate risks of over-reliance or uncritical acceptance of AI-generated content. Understanding these interaction effects is crucial for developing responsible AI usage guidelines and training programs tailored to different user profiles.
IntroductionGenerative artificial intelligence (GAI) is increasingly integrated into daily decision-making and task performance, as users rely on it to improve productivity, support decision-making, and reduce perceived human error. However, GAI outputs may contain inaccuracies, misleading content, or algorithmic biases. These issues may contribute to uncritical acceptance or excessive dependence among users, potentially posing risks to individuals and third parties. From an individual perspective, this study systematically explores how personality traits and text-based GAI usage experience interact in relation to users’ trust in and dependence on text-based GAI.MethodsThis study was based on a survey of 502 users measuring personality, usage experience, trust, dependence, self-efficacy, and critical thinking.ResultsSignificant interaction effects were found. Among high-experience users, the Well-Adjusted Type showed the highest dependence and self-efficacy, along with relatively high trust and critical thinking. The Outgoing-Unstructured Type showed the highest overall trust and, among high-experience users, the highest critical thinking, while its dependence did not differ significantly between low- and high-experience users. Overall, the Emotionally-Sensitive Type showed the lowest mean levels across the four outcomes, which may indicate a more cautious pattern of text-based GAI engagement.DiscussionThese findings are discussed in relation to an evolutionary-cognitive perspective on how users may engage with human-like communicative cues in text-based GAI. The study emphasizes the importance of personalized text-based GAI usage strategies to mitigate misuse risks.