Personality Traits as Predictors for AI Use Among Gen Z: Performance Tool vs. Social Agent
Alexandra Fisher
Department of Psychology, University of Pennsylvania
PSYC 4450: Advanced Research Topics in Technology and The Good Life
Dr. Angela Duckworth & Dr. Lyle Ungar
Overview
Artificial intelligence has become deeply woven into the everyday lives of Generation Z, but this research argues that young people are not all using AI for the same reasons. Using survey data from over 2,300 Gen Z participants across the United States, this study explored how personality traits and emotional experiences shape the way young people interact with AI. Specifically, it examined whether achievement-oriented self-beliefs, social competence, and loneliness predict two distinct forms of AI use: productivity use (using AI for tasks like writing, homework, research, or problem solving) and social-agent use (using AI as a conversational companion, confidant, or source of personal advice). The findings revealed that highly motivated and achievement-oriented individuals were more likely to use AI as a productivity tool, helping them work faster, organize ideas, and pursue goals more efficiently. In contrast, loneliness emerged as the strongest predictor of social AI use, suggesting that many young people turn to AI not simply for entertainment, but because they are seeking connection, support, or someone to talk to. Importantly, the study found a “double dissociation,” meaning the psychological factors predicting productivity AI use were different from those predicting social AI use. This suggests that AI is not functioning as one single technology in Gen Z’s lives, but instead as a flexible resource fulfilling very different human needs depending on the individual using it. Ultimately, the research argues that conversations about AI should move beyond simply asking whether young people use AI and instead focus on why they use it, what emotional or motivational needs it may be fulfilling, and what this reveals about the social realities of growing up in a generation where AI is a normal part of daily life.
Controlling for age, gender, and household income, those with stronger achievement-oriented beliefs were significantly more likely to use AI for productivity-related tasks (r = .090, p < .001), consistent with H1. In other words, individuals who are motivated to succeed and set high goals appear to adopt AI in ways that serve those goals, using it to support their work, writing, and academic tasks. That said, the effect was modest, suggesting that achievement orientation amplifies productivity AI use rather than determining it. Gen Z uses AI for tasks broadly, but ambition tilts the needle slightly further in that direction.
What is more striking is what achievement beliefs did not predict. Highly ambitious individuals showed no tendency to use AI as a social agent, with the association being essentially zero (r = −.006, ns), and Steiger’s test confirmed that the two correlations are significantly different from one another (Δr = .096, p < .001). Highly ambitious individuals were no more and no less likely to use AI as a friend, conversational partner, or source of personal advice. Together, these findings suggest that ambition selectively predicts how individuals use AI as a tool for getting things done, but not as a source of interaction or connection.
Loneliness emerged as one of the most powerful predictors of how individuals engage with AI, but not all forms of use were equally affected. Lonelier individuals were somewhat more likely to use AI for productivity-related tasks (r = .085, p < .001), but this effect was modest. A possible explanation is that loneliness may reflect a broader tendency to turn to available tools and resources, rather than a specific drive toward productivity itself.
However, the picture changed dramatically for social agent use. Lonelier individuals were substantially more likely to use AI as a friend, conversational partner, or source of personal advice (r = .260, p < .001), creating the largest single predictor-outcome relationship in the study, and consistent with H2. Steiger’s test confirmed that this difference in correlations was not incidental: loneliness was significantly more strongly associated with social than productivity AI use (Δr = .175, p < .001).
What these findings suggest is that lonelier individuals are not turning to AI to get more done; they are turning to it for something closer to connection, orienting them toward relational, socially substitutive forms of interaction, a distinction that carries real implications for how we understand the role AI is beginning to play in the social lives of young people.
Individuals with higher social competence showed a distinct and balanced pattern of AI engagement. Those who are more socially capable were significantly more likely to use AI for productivity-related tasks (r = .122, p < .001), consistent with H3. This pattern suggests that social competence may reflect a broader sense of self-efficacy, extending beyond interpersonal contexts to include the adoption and effective use of new technologies. Socially capable individuals appear to engage more actively with AI as a tool, much as they might engage more actively with other resources available to them.
At the same time, social competence also predicted social agent use (r = .075, p < .001), though the relationship was weaker. Though this may seem counterintuitive, the findings suggest that it is not only socially anxious, lonely, or avoidant people who are turning to AI for conversation and connection, but also those who are already comfortable around people. The more plausible explanation is that, while socially confident individuals may not prefer AI companionship, their personalities reflect a general willingness to engage with both people and technology.
Steiger’s test confirmed that social competence was more strongly associated with productivity than social AI use (Δr = .047, p = .013). This suggests that, unlike loneliness, which shows a sharp and targeted link to social AI use, social competence is associated with a more general increase in AI engagement across contexts, with only a slight tilt toward productivity.
Why This Matters
Gen Z are the first generation to grow up with AI as a normal part of daily life. But what this research reveals is that they are not using it as one thing. Some young people use AI to get things done, to write essays, solve problems, and power through their work. Others use it for something entirely different: to talk, to feel heard, to get advice about their relationships and their lives. And what determines which? It turns out to be less about the technology and more about the person using it.
Young people who are motivated, goal-oriented, and driven to succeed are more likely to reach for AI as a tool. That makes intuitive sense. What is more surprising is what predicts social AI use. It is not shyness. It is not a preference for screens over people. It is loneliness. The young people most likely to be talking to AI as a friend are those who feel genuinely disconnected from the people around them. They are not avoiding human connection. They are missing it.
That finding reframes the conversation about AI in young people’s lives. The question is not whether Gen Z is spending too much time with AI. The question is what that time is telling us about how connected, or disconnected, they feel. AI is not the story here. Loneliness is. AI is simply where we can see it.
For anyone building AI products, making school policy, or raising or working with young people, this matters. Productive AI use is already widespread across this generation and is unlikely to slow down, the more useful question is how to support it thoughtfully. But social AI use is a signal worth paying attention to. When a young person is turning to a chatbot for companionship and advice, that is not just a technology behavior. It is a window into their social world, and it deserves a response that goes beyond the screen.