HeartBot Trial

Can AI conversation encourage women to make life-saving decisions before a heart attack happens?

The HeartBot Trial studies how an AI-powered conversational chatbot can improve heart attack awareness, health knowledge, prevention attitudes, self-efficacy, and intentions to adopt heart-healthy behaviors. Through this multi-year project, the CHATR Lab investigates how an expert team-designed AI chatbot can deliver personalized health communication and promote meaningful heart health literacy. Early studies have shown that interacting with HeartBot significantly improves women's awareness of heart attack symptoms and appropriate care-seeking behaviors.

Over the past three years, the project has been supported by three internal research grants from the University of California San Francisco (UCSF) and has generated multiple peer-reviewed publications.

Team. The HeartBot Trial is a collaborative effort led by Prof. Yoshimi Fukuoka at UCSF, with the CHATR Lab contributing to the design and evaluation of pilot tests and clinical trials.

Selected Publications

  • Suzuki, H., Zhang, J., Kim, D. D., Sagae, K., DeVon, H. A., & Fukuoka, Y. (2026). Message humanness as a predictor of AI’s perception as human: secondary data analysis of the HeartBot study. JMIR AI, 5(1), e67717. doi: 10.2196/67717
  • Kim, D. D., Zhang, J., Sagae, K., Devon, H. A., Hoffmann, T. J., Rountree, L., & Fukuoka, Y. (2025). Human-delivered conversation versus AI chatbot conversation in increasing heart attack knowledge in women in the United States: quasi-experimental studies. Journal of Medical Internet Research, 27, e73184. doi:10.2196/73184
  • Fukuoka, Y., Kim, D. D., Zhang, J., Hoffmann, T. J., DeVon, H. A., & Sagae, K. (2025). AI HeartBot to increase women’s awareness and knowledge of heart attacks: nonrandomized, quasi-experimental study. JMIR Cardio, 9(1), e80407. doi: 10.2196/80407

Persuasive AI Index

Can AI accurately judge what humans find persuasive?

As AI-generated persuasive messages become increasingly common and infused into the current digital information ecosystem, understanding whether AI can also reliably evaluate their effectiveness is both an interesting question and an important challenge. Through the SabotageIndex project, the CHATR Lab is conducting human-subject research comparing how people evaluate AI-written and human-written counterarguments across all topics (social, political, cultural, scientific, legal, etc.). These human judgments will be compared with an AI judge model trained to assess the persuasive strength of messages produced by humans and large language models.

Our goal is to identify where AI and human evaluations align—and where they differ—to improve the reliability and safety of AI systems used to generate and evaluate persuasive communication. Data collection is currently underway for Summer 2026.

Team. Funded by a $1 million Coefficient Giving grant, the larger project is led by Prof. Weiyan Shi (PI) from Northeastern University. CHATR Lab is conducting the human evaluation research.

Using Positive Friction to Reclaim Human Agency in the AI News Environment

Can slowing down our interactions with AI make us better (critical) news consumers?

As generative AI makes accessing and reading news faster and easier, it may also encourage passive information consumption and reduce critical thinking. This project investigates whether introducing "positive friction"—small moments that prompt users to pause and reflect—can help people engage more thoughtfully with AI-generated news and strengthen their sense of agency.

The research combines interviews with young adults and a longitudinal field experiment to test how positive friction interventions influence news evaluation, critical thinking, and human-AI interaction over time. The findings will inform the design of more human-centered AI systems and provide practical guidance for educators, designers, and the public.

Team. This project is led by Prof. Kecheng Fang from The Chinese University of Hong Kong and is supported by a General Research Fund from Hong Kong (HKD 639,920). Dr. Jingwen Zhang serves as a Co-Investigator. The project began in summer 2026.