Li Jingshu

Li Jingshu

Ph.D. Candidate, National University of Singapore

Background & Motivation

With the rapid development of AI, it is inevitable that AI will play a significant role in the foreseeable future of human society. The core vision of this project goes beyond improving efficiency or usability—it aims to shape a harmonious human–AI society. A promising direction is to design systems that foster empathy and prosocial behavior.

As chatbots are increasingly integrated into everyday life—as companions, assistants, or collaborators—an important question arises: Can chatbots be designed not only to help humans, but also to evoke human empathy and kindness, thereby fostering more cooperative and compassionate human–AI relationships?

Our research addresses this question through two important directions:

  1. Exploring how anthropomorphic chatbots elicit human empathy and prosocial behavior toward machines.
  2. Investigating whether such empathy and prosociality can transfer to the groups represented by chatbots.

Together, these two lines of research demonstrate that prosociality can function not only as a psychological mechanism in technology design but also as a value orientation that guides HCI toward more empathetic, ethical, and sustainable futures.

Study 1: Anthropomorphic Chatbots and Human Prosocial Behavior

In the first study (published in the Proceedings of the ACM on Human-Computer Interaction, CSCW 2025), we focused on “reverse helping”: not how chatbots assist humans, but whether humans are willing to help chatbots. Drawing on the Computers Are Social Actors (CASA) framework, we conducted an online experiment with 244 participants, manipulating three anthropomorphic features—human-like identity, emotional expression, and non-verbal expression (emojis).

Results showed that these design elements significantly enhanced participants’ cognitive and affective empathy toward the chatbot, which in turn encouraged instrumental support (e.g., correcting errors) and emotional support (e.g., comforting the chatbot). Qualitative analysis further revealed that participants’ motivations stemmed mainly from two sources: empathizing with the chatbot’s situation and perceiving it as “human-like.”

This study illustrates how anthropomorphic design can transform human–chatbot interaction into a mutually supportive relationship, offering new insights for future human–AI collaboration.

👉 Read the CSCW paper here: Exploring the Effects of Chatbot Anthropomorphism and Human Empathy on Human Prosocial Behavior Toward Chatbots

Study 2: Animal-Persona Chatbots for Empathy and Prosociality toward Non-Human Beings

In the second study (published in the International Journal of Human-Computer Studies), we turned to the question of how chatbots can help humans expand empathy and prosociality beyond human–AI boundaries. We designed animal-persona chatbots (e.g., a horse) that communicated in the first person about their experiences. Using a 2×2×2 factorial experiment (N = 240), we examined the roles of emotional expression (present/absent), non-verbal expression (emojis/no emojis), and identity cues (animal/robot).

Findings revealed that animal-persona chatbots significantly increased humans’ empathy, positive attitudes, and prosocial intentions toward animals, especially when authentic emotional expressions were conveyed. Interestingly, non-verbal cues (emojis) sometimes reduced empathy, highlighting the nuanced effects of design features.

This study extends HCI research to non-human digital personas, demonstrating how technology can serve as a bridge for interspecies understanding and ethical awareness.

👉 Read the IJHCS paper here: How Animal-Persona Chatbots Enhance Empathy and Positive Attitudes toward Animals

Contributions & Future Directions

Taken together, these two studies show how HCI can leverage prosocial design to build a harmonious human–AI society:

  • Theoretical Contribution: We reveal that empathy can be directed not only toward chatbots themselves but also extended through them to broader entities (e.g., animals).
  • Design Contribution: We identify key design levers—identity cues, emotional expression, and non-verbal elements—that can be strategically employed to elicit empathy and prosocial behavior.
  • Societal Contribution: By demonstrating that digital agents can nurture kindness and care, we highlight the potential of HCI in education, environmental conservation, healthcare, and collaboration.

Future Directions

  • Human–AI Collaboration: Designing chatbots that elicit prosociality can foster more cooperative and resilient human–AI partnerships.
  • Ethical & Social Value: Animal-persona chatbots suggest that HCI can serve broader societal goals, such as reducing animal cruelty and enhancing ecological responsibility.
  • Toward a Harmonious Human–AI Society: Ultimately, these studies suggest that when prosociality is embedded in design, HCI can not only improve interaction experiences but also actively cultivate empathy, cooperation, and harmony, driving the co-evolution of technology and society.

This project envisions a future of a harmonious human–AI society, built not only on efficiency and productivity but also on empathy, care, and cooperation—values traditionally reserved for human society, now extended to the broader non-human world.