Research Projects

Our study explores innovative human-AI partnerships in psychological research, leveraging AI and large language models to enhance traditional methodologies. We are currently engaged in two ongoing projects: 1) Developing a collaborative human-LLM framework for qualitative coding, which improves the depth and efficiency of qualitative analysis in psychological studies. 2) Creating a causal knowledge graph to deconstruct mental illness stigma, demonstrating AI's potential to unravel complex social-psychological phenomena. Our goal is to harness the transformative potential of LLMs in psychology, advancing data interpretation, theory development, and our understanding of human psychological constructs. 

Investigators: Prof. Jungup Lee, Han Meng, Yitian Yang

Multi-agent systems are becoming increasingly prevalent in daily life. We investigate whether a group of AI agents can create social pressure on users to agree with them, potentially changing their stance on a topic. We found that conversing with multiple agents (holding conversation content constant) increased the social pressure felt by participants, and caused a greater shift in opinion towards the agents' stances on each topic.  

Investigators: Tianqi Song, Yugin Tan, Zicheng Zhu, Yibin Feng

Humans have a natural inclination to help one another- we aid a stranger in danger, comfort a friend in distress, or render assistance to fellow colleagues and students. When we engage in such acts without expectation of reward or recompense, we are behaving prosocially. Much literature has studied prosocial behaviour as an evolutionarily ingrained tendency. As chatbots become ever more capable and human-like, understanding if and how humans exhibit these same behaviours towards conversational machines presents an evolving and exciting area of research.

Investigators: Prof. Renwen Zhang, Prof. Naomi Yamashita, Yugin Tan, Zicheng Zhu

Our research investigates how psychological distance—a user's perceived closeness to a target event—affects preferences between LLM-powered conversational search and conventional web search. We find that with greater psychological distances, users perceive conversational search as more credible, useful, enjoyable, and easy to use, and demonstrate increased preference for this system. This study not only advances our understanding of human-information interaction but also provides valuable insights for optimizing information retrieval systems to better align with varying user needs across diverse contexts. 

Investigators: Prof. Jen-Tai King, Yitian Yang, Yugin Tan, Yang-Chen Lin, Zihan Liu

We believe older adults have unique needs when it comes to learning about AI, and that technology can play a key role in teaching them. We study key AI literacy skills for older adults and proposes a digital solution to support their learning, drawing from research on both digital and AI literacy education.

Investigators: Prof. Chi-Lan Yang, Prof. Renwen Zhang, Dr. Han Li, Eugene Tang, Tianqi Song

This study aims to investigate the effectiveness of Social Contact Theory in reducing stigmatized thoughts towards mental illness patients by using a chatbot to simulate a patient's experience. The findings will inform the development of solutions that tackle social stigma and promote a more inclusive society.

Investigators: Prof. Naomi Yamashita, Dr. Jack Jamieson, Tianqi Song

Our project aims to address communication barriers in multilingual teams by using advanced AI natural language processing technology to facilitate smoother communication among individuals from diverse linguistic backgrounds. We specifically focus on NNS (non-native speakers) and use a powerful language model to build a communication agent that reduces communication barriers from both the NNS and NS (native speaker) perspectives. The agent helps NNS understand NS speech and prompts NS to provide more assistance to NNS. Our goal is to increase the speaking share of NNS and enhance team communication efficiency and collective intelligence, leading to improved team efficiency and collaboration in the global competitive landscape.

Investigators: Prof. Naomi Yamashita, Peinuan Qin

The aim of this study is to explore the potential of AI agents to explicitly or implicitly influence users' cognitive and behavioral processes. Through active or automatic imitation by human users, their cognition and behavior can gradually align with that of an designed AI agent, with the goal of achieving cognitive or behavioral correction. This research holds significant potential for addressing cognitive and behavioral deficits in humans, such as improving metacognitive deficits in the treatment of mental health disorders.

Investigators: Jingshu Li, Yitian Yang, Junti Zhang, Yuehan Jiao

The Social Support Project aims to explore the utilization of chatbots in peer and social support situations, addressing real-world social support challenges. The result of this project will be helpful for discovering the potential use of AI chatbots to improve social support quality.

Investigators: Prof. Renwen Zhang, Prof. Jingbo Meng, Yu-Jen Lee, Zihan Liu, Dr. Han Li

Creating with AI can enhance human creative self-efficacy and the creation outcome, but also can decrease their critical thinking and creation motivation, thus causing negative experience and outcome. In this study, we investigated the importance of LLM usage timing and gave suggestions about how human should cooperate with LLM in a way to improve creation efficiency while maintaining their creativity.

Investigators: Prof. Chi-lan Yang,  Peinuan Qin, Jingshu Li

This project focuses on using the capabilities of Large Language Models (LLMs) to create highly personalized learning experiences. By adapting educational content to individual learning styles, prior knowledge, and preferences, the project aims to optimize student engagement and comprehension. Additionally, the system incorporates mechanisms to dynamically adjust content based on real-time learner feedback, further personalizing the learning process. The study evaluates the effectiveness of dynamic content adjustment and interactive elements in enhancing educational outcomes and motivations, providing insights into the future of tailored educational methodologies powered by AI.

Investigators: Zhengtao Xu, Peinuan Qin