Report on the 27th B’AI Book Club (※Research Presentation)
“Exploratory analysis of perceptions of artificial intelligence among high school students living in Japan”

On February 27th, 2024, a hybrid meeting was hosted for the B’AI Global forum’s 27th Book Club. The meeting featured two of the graduate student members, William Guzman and Alyssa Castillo Yap, presenting on their ongoing research regarding the perceptions of AI among high school students in Japan. 

Will and Alyssa presented a brief literature review on the topic as well as a review of data collection and ongoing analysis. Many studies examining students’ relations with AI currently seek to “correct” students’ “misconceptions” about AI in accordance with recently standardized educational frameworks (Mertala et al 2022; Mertala and Fagerlund 2024). Few studies, however, have explored the reasons behind why or how students form their perceptions of AI in the first place, and which forms and examples of AI may have led to the formation of their opinions. Additionally, many studies continue to focus on American and European contexts, despite significant research demonstrating the diversity of opinions towards AI found in different countries and cultural contexts. To this end, this study investigates the means of exposure to and general perceptions of AI among high school students living in Japan.

The researchers presented their methods which included a bilingual, online survey of 86 high school students living in Japan. Through a mix of open-ended and Likert-scale questions, researchers sought to understand (a) the general perceptions of AI among the students and (b) their main forms of exposure to different AI technologies to determine how that impacted their opinion formation. Due to the complexity of the open-ended questions and responses, the researchers sought feedback on the best ways to characterize the data, and many helpful suggestions were offered around the ways deductive, inductive, and abductive coding schema could be applied to explore the qualitative elements of the open-ended questions. Particular interest was taken in the potential range of responses from single participants. Book club members were interested in variability presented by a single respondent and the ways perceptions of AI might vary based on the context or topic at hand. Additional insight was offered on the structure of Japanese secondary education, particularly focused on the ways students are taught to respond to certain question types. This prior training and the different approaches students have to answering questions could be a confounding variable in the analysis, but also can be helpful for developing clear insights from the responses.

The quantitative data were also discussed. They are in some ways open to more streamlined forms of analysis via standard statistical methods which were discussed simultaneously as a fruitful avenue for further research and a useful means to complicate and deepen qualitative analysis. Book club attendees specifically explored questions of how Likert-scale responses could be coordinated with qualitative analysis and the ways we could shed light on the diversity and complexity of perspectives on AI. From these discussions, the researchers gained many useful suggestions and insights into next steps in the research and data analysis process.

The tentative results of the study indicate two main correlations: that between a student’s gender and their likelihood of being exposed to specific technologies and the impact that specific forms of exposure, for instance ChatGPT, has on perceptions of AI as dangerous, ubiquitous, useful, and accessible. Participants of the book club found these insights very enlightening and further discussed the ways that this study and potential future work could contribute to ongoing discussions around AI ethics, education, and advocacy.

This research was presented as a continuation of the B’AI Global Forum graduate student members’ exploration of artificial intelligence following the exhibition entitled “Perceptions: Reconstructing AI Narratives” in September of 2023. The graduate student members are currently working to understand various possible and valid “perceptions” about AI that people in different socio-cultural contexts may hold. Interactions with AI need not only be a matter of cognitive interpretation and definition, but should also be situated within the time and space of a user’s and designer’s realities. There is a need for more open-minded interpretations of AI spanning diverse use cases and applications across different socio-cultural contexts. Academic research approaches should allow for expansive understandings of human-nonhuman interactions with the advent of AI in daily use, and seek to debunk overly apocalyptic or prescriptive definitions of AI in non-Western settings.

The members greatly appreciated all of the comments received from the book club session and are continuing their investigation over the course of the spring.