Report on the 26th B’AI Book Club (※Research Presentation)
“Survey on the Awareness of AI/Algorithms and Gender Equality Among Teenage and Twenties Women”

On January 23, 2024, the 26th session of the “B’AI Book Club” was held in a hybrid format, featuring an interim report from the “AI Algorithms and Gender Inequality” research group, one of the projects under the B’AI Global Forum.

This research aims to illuminate the relationship between AI/algorithms and gender inequality by examining how young women (aged 18 to 29), a significant user group of digital technology embedded with artificial intelligence (AI) and algorithms, perceive the influence of this relationship on their own perceptions and choices in contemporary society. The study investigates two main aspects: (1) the negative impact of the relationship between young women and “recommender systems” on their “confidence,” “self-esteem,” and “self-assessment”; and (2) the actual situation regarding the changes in values due to young women’s dependence on “self-management apps” and the management/control of their lives/gender by the apps (AI/algorithms). The research employed online questionnaire surveys and interviews. The initial phase involved a questionnaire survey, which included both multiple-choice and free-response questions, to understand usage patterns. This was followed by in-depth interviews with 20 participants.

At the interim stage, tentative conclusions can be summarized into three main points. Firstly, there is a mixed awareness regarding advertising and recommender systems embedded with AI/algorithms. Interviewees expressed convenience in seeing “contents similar to their interests displayed” based on past search history and input data, while also harboring concerns about the collection and potential misuse of personal information. However, they acknowledged limited means of resistance and found themselves somewhat resigned to the situation, accepting it as somewhat inevitable. Also, it was found that some interviewees actively managed their settings to prevent excessive data collection or attempted to resist as much as possible.

Secondly, there is a noted lack of awareness among users regarding data collection by menstrual cycle tracking apps. These apps guide and support users’ lives by visualizing their physical conditions as objective data. Regarding how users engage with these tools, they largely fall into three patterns: (1) attempting to explain physical conditions from the data, (2) feeling disappointment, confusion, or anxiety, and (3) perceiving it indifferently as mere recording. Many interviewees acknowledged the risk of “menstrual cycle” information leaking alongside personal data through data input. However, it became evident that some participants, considering the information less significant, were unconcerned about potential leaks.

Thirdly, the relationship between social media usage and gender norms is highlighted. In today’s social media landscape, there is an inundation of images depicting women as “thin, cute, and beautiful,” propagated by celebrities and influencers. Even among “ordinary people,” there is a tendency to conform to these feminine gender norms. From the narratives of interviewees, it became apparent that despite holding a critical stance towards social media, individuals could inadvertently become ensnared in lookism or internalize self-blame. Moreover, concerns were expressed regarding younger generations, who are deeply immersed in social media from an early age, being particularly vulnerable to these norms and experiencing significant negative impacts as a consequence.

One limitation of this research is its reliance on snowball sampling to recruit interview participants, leading to a slight bias towards individuals with “higher education” and a skewed representation of the “middle class” and “cisgender women.” Consequently, it is important to acknowledge that the response tendencies observed in this survey may not fully reflect the broader societal context.

In the discussion following the presentation, participants raised several important points. Firstly, they suggested that further analysis or inquiries might be necessary to understand the lack of awareness regarding the collection of data by menstrual cycle tracking apps, unlike with recommender systems or lookism. Secondly, there was a discussion on addressing the assumption of gender social norms in the responses and narratives of research participants. Participants also debated the necessity of conducting awareness surveys regarding seemingly gender- or femininity-unrelated information handled on systems supported by similar AI/algorithms, given that all the topics being dealt with are overtly linked to gender or femininity. Particularly, there was a suggestion to consider how gender binary concepts, especially reflected in advertisements and recommender systems, are positioned within the context of Japan, and how discussions can be compared with research conducted outside Japan. These issues are essential points for future consideration.

The session provided valuable insights into how various digital technologies, pervasive in everyday life and based on AI/algorithms, shape, perpetuate, and reproduce gender inequality. It also delved into the intertwining of these technologies with the self-perception of female users, prompting considerations for future research directions.