“`html
Artificial intelligence is often promoted as society’s ultimate equalizer—revolutionizing businesses, optimizing tasks, and enhancing lives for all—but for numerous marginalized Americans, AI seems far from a promise.
Instead, it appears as a peril.
A recent analysis from the University of Michigan indicates that gender minorities, women, and individuals with disabilities—especially those who are neurodivergent or dealing with mental health challenges—exhibit significantly more unfavorable attitudes towards AI compared to their majority-group counterparts.
Utilizing data from a nationally representative survey involving 742 individuals, researchers solicited participants’ views regarding AI’s influence on their lives and occupations. The results showcase a pronounced disparity in positivity towards AI based on identity factors.
“AI may be ubiquitous, yet it’s not for everyone—at least, not yet,” stated lead author Oliver Haimson, assistant professor at the University of Michigan School of Information and Digital Studies Institute. “If we disregard the perspectives of marginalized individuals, we risk constructing an AI-driven future that exacerbates inequalities rather than alleviating them.”
Major insights:
- Nonbinary and transgender respondents expressed the most critical attitudes towards AI overall.
- Women were significantly less inclined than men to have a favorable outlook on AI.
- Disabled participants, notably those who are neurodivergent or struggling with mental health issues, also reported negative attitudes toward AI.
- Interestingly, Black participants showcased more favorable views of AI compared to white participants, indicating a more complex interaction with technology among people of color.
What contributes to the skepticism?
The study underscores tangible harms that may lead to negative perceptions of AI:
- Facial recognition technology that misgenders or inaccurately categorizes transgender and nonbinary individuals—frequently while monitoring them.
- Predictive policing algorithms that reinforce racial biases, resulting in unjust detentions.
- Healthcare systems dependent on AI frameworks not specifically designed for disabled individuals.
“These are not hypothetical issues,” Haimson remarked. “Individuals who are apprehensive about AI frequently have lived experiences with systems that misidentify, exclude, or harm them due to their gender identity or disabilities.”
Why this is significant
The investigation challenges the prevailing narrative of AI as a neutral or universally advantageous instrument. It calls on technologists, corporations, and policymakers to reconsider and question: Who benefits from this technology—and who is left behind?
The study’s co-authors include doctoral candidate Samuel Reiji Mayworm, research fellow Alexis Shore Ingber, and assistant professor Nazanin Andalibi.
This research was presented at the recent 2025 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’25)—a premier event for discussions at the crossroads of AI and ethics.
“`