I study the trajectory of AI in providing social and emotional support. My research investigates two central questions: (1) What types of social support do humans seek from AI? and (2) What types of support are most effective when delivered by AI?

​

As part of my research, I use Qualtrics to embed AI chatbots, enabling real-time human-AI interactions within experimental and field study contexts. I’ve made the underlying code available through an open GitHub repository to support other social science researchers in using this method in their own work. In the future, I plan to develop a more comprehensive website that offers greater flexibility in customizing interaction styles, prompt design, and chatbot settings within Qualtrics, and will update the repository accordingly.

​

In one line of work, I examine the kinds of support people actively seek from AI companions. For instance, in one study, we explore the real-life social interactions of individuals who rely on AI companion apps. In another project, we analyze real-world AI therapy sessions conducted in workplace settings to understand the kinds of emotional and psychological support employees turn to AI for. Another study investigates people’s expectations of gender congruence in AI assistants, examining whether mismatches with this expectation reinforce gender stereotypes.

​

In a second line of research, I examine the effectiveness of AI in delivering social and emotional interventions. In one study, we assess whether emotion regulation interventions from AI can help employees cope with workplace anger, and whether strategies to enhance perceived AI empathy improve outcomes. In another large-scale project, we deliver AI-facilitated behavioral interventions (e.g., social belonging, values affirmation, growth mindset, cognitive reappraisal) to students across more than 20 U.S. colleges. These interventions are administered either individually or with the AI integrated as a group member, allowing us to test the comparative effectiveness of different delivery formats on students’ academic and social outcomes. I also study how persuasive strategies used by AI can reduce belief in conspiracy theories. 

 

Under the field of AI providing social and emotional support, I'm also interested in studying how people respond differently when they realize that support came from an AI rather than a human—a phenomenon known as AI aversion. I investigate whether increasing perceived AI empathy and expertise can mitigate this aversion.

​

Beyond my work on AI, I am also broadly interested in emotional processes, including loneliness and the use of cognitive reappraisal as an emotion regulation and general reframing strategy.