Graduation Year

2026

Document Type

Open Access Senior Thesis

Degree Name

Bachelor of Arts

Department

Psychology

Reader 1

Stacey Wood

Reader 2

Jose Arreola

Abstract

The proposed study is intended to explore the impacts of artificial intelligence (AI) tools when used in cognitive behavioral therapy (CBT) to treat anxiety in American adults. The research is designed to measure whether AI systems such as large language models (LLMs) provide sufficient therapy through the factors of empathy, coping skills, and anxiety symptom reduction. Human therapists are hypothesized to surpass chatbots in all three. Through an experimental design, participants will be placed in either an AI therapy condition or human therapy condition. Participants will undergo a 12-week CBT program, after which they will measure anxiety reduction, perceived empathy of their therapist, and the efficacy of coping skills. It is anticipated that AI therapists will underperform human therapists in all three tested hypotheses, with participants in the human condition reporting lower anxiety, higher empathy, and better coping skills as a result of their therapy. This research has significant implications on the future of mental health support, as AI therapy is becoming more prevalent and advertising proper CBT. Shedding light on such novel technology is urgent for improving mental health resources in the United States.

Download

Share

COinS