Understanding AI Perceptions Among Marginalized Groups

A recent study led by Dr. Oliver Haimson and his colleagues at the University of Michigan has revealed significant skepticism towards artificial intelligence (AI) among marginalized populations, particularly nonbinary, transgender, and disabled individuals. Initially focused on augmented reality, the research pivoted to examine AI perceptions after participants expressed concerns about the technology during data collection. The findings, published in June in the article “AI Attitudes Among Marginalized Populations in the U.S.: Nonbinary, Transgender, and Disabled Individuals Report More Negative AI Attitudes,” indicate that nonbinary individuals exhibit the most negative views towards AI, largely due to their rejection of categorization, which contrasts with AI's reliance on categorizing data.

The Advocate

The study highlights that multiracial participants also expressed skepticism, as did individuals with disabilities, particularly those identifying as neurodivergent or with mental health conditions. Transgender individuals reported more negative attitudes towards AI compared to their cisgender counterparts, and women generally demonstrated lower trust in AI than men. The researchers noted that AI can cause harm to gender minorities through issues such as algorithmic misgendering and violations of privacy and consent.

The Advocate

A critical concern raised by Haimson is the issue of consent regarding AI usage. He pointed out that individuals often do not have the opportunity to opt-in or opt-out of AI applications in various contexts, including employment and healthcare. “Sometimes [AI is] being used on you without your permission or consent,” Haimson stated, emphasizing the lack of transparency surrounding data usage. This lack of clarity contributes to the distrust felt by marginalized groups.

The Advocate

The study underscores the gap in AI development concerning the needs of trans and nonbinary individuals. Haimson noted that many companies creating AI technologies often overlook these communities. In response, trans digital creators are developing technologies that cater specifically to their needs. An example provided is the ShotTraX app, designed to help users track hormone shot doses for gender-affirming healthcare. Haimson remarked, “There’s so many mainstream technologies that just don’t really think about trans users and nonbinary users,” highlighting the importance of agency in technology creation for marginalized groups.

The Advocate

While the study's findings were not surprising to Haimson and his team, they provide empirical evidence that can inform future research and development in AI technologies. Understanding the unique perspectives of marginalized populations is essential for creating inclusive and effective AI systems that respect the rights and identities of all users.

The Advocate

• “If you think about nonbinary identity, it’s really about rejecting categorization and rejecting being put into these boxes,” — Dr. Oliver Haimson, Associate Professor of Information

The Advocate

• “The issue of consent is also of great concern, he says “Sometimes [AI is] being used on you without your permission or consent.” — Dr. Oliver Haimson

The Advocate

• “There’s so many mainstream technologies that just don’t really think about trans users and nonbinary users,” — Dr. Oliver Haimson

The Advocate

This research highlights the need for greater inclusivity in AI development, ensuring that the voices and concerns of marginalized communities are acknowledged and addressed.

The Advocate