Last week, Prof. Dr. Kaiping Chen was a guest at DIID. During a workshop, some DIID members had the opportunity to exchange ideas with the American colleague. In the evening, Dr. Chen gave a lecture as part of the HeiCAD Lectures.
On March 15, 2023, Prof. Dr. Kaiping Chen (Assistant Professor in Computational Communication at the Department of Life Sciences Communication at the University of Wisconsin-Madison) was a guest at the Düsseldorf Institute for Internet and Democracy (DIID). In the morning, some members of DIID had the opportunity to exchange ideas with their colleague, who conducts research on topics including digital inequality, online deliberation, and human-computer interaction, during a research workshop. During the workshop, DIID researchers presented current and completed research projects in the AI context, identifying significant intersections with Dr. Chen’s research and discussing collaboration options.
In the evening, Kaiping Chen gave a lecture as part of the HeiCAD Lectures entitled How GPT-3 responds to different publics on climate change and BLM (more information). After a brief welcome by DIID coordinator Dr. Dennis Frieß, Dr. Chen presented a study in which subjects engaged in a dialogue with the chatbot GPT-3 – the predecessor model of the currently much-discussed chat GPT. The study focused on the user experience and opinion changes of people who do not consider climate change to be real or who are critical of the Black Lives Matter movement (opinion minorities). This found a significantly poorer user experience with GPT-3 within the opinion and education minorities. However, at the same time, these two groups also gained the most knowledge and changed their attitudes after the chat regarding the Black Lives Matter movement and the existence of climate change. The differences in user experience were primarily attributed to differences in GPT-3’s conversational style. For example, the chatbot used more negative language when interacting with representatives of minority opinion and education than with representatives of majority opinion. During her presentation, Dr. Chen discussed the implications of the findings for a deliberative AI system that focuses on diversity, equity, and inclusion. A pre-published version of the study can be found here.
We thank Dr. Kaiping Chen for the exciting talk and academic exchange. To all participants, we thank you for the input provided and the stimulating discussion. Finally, we would like to thank HeiCAD, especially Saskia Reither and Martin Mauve, who made Kaiping Chen’s invitation possible.