A new study, which draws on survey data from approximately 3,000 respondents in the United States and Canada, finds that women, on average, exhibit greater skepticism toward artificial intelligence (AI) technologies than men. Analysis suggests that differences in risk orientation and exposure help explain this gender gap.
A paper describing the research appears in the journal PNAS Nexus on January 20. Caltech's R. Michael Alvarez, the Flintridge Foundation Professor of Political and Computational Social Science, is a co-author of the study.
"In earlier work, we found that women are consistently more skeptical of AI than men, but we couldn't fully explain why. This study shows that risk is a key part of the puzzle," says former Caltech postdoc Beatrice Magistro, a study co-author who is now an assistant professor at Northeastern University. "These attitudes matter because they translate into real differences-in who adopts AI, who shapes its development, and what kinds of policies people support. Whether or not AI delivers on its promises, these gaps are already creating new axes of inequality."
The data were collected in November 2023 through an opt-in panel organized by YouGov.
According to the research team, understanding these gender-specific perspectives is essential for designing AI policies that promote equitable outcomes and avoid reinforcing existing inequalities.
A significant portion of this work was supported by research incubator funding from the Caltech Linde Center for Science, Society, and Policy (LCSSP) during Magistro's tenure as a postdoctoral scholar research associate in computational social science at Caltech. Alvarez is co-director of the LCSSP, which also provided funding for in-person workshops for the research team and supported a conference on the political economy of AI , at which the work was discussed.
The paper is titled "Explaining Women's Skepticism toward Artificial Intelligence: The Role of Risk Orientation and Risk Exposure." In addition to Magistro and Alvarez, co-authors are Sophie Borwein, Bart Bonikowski, and Peter J. Loewen. Additional funding for the work was provided by the Social Sciences and Humanities Council of Canada, through a grant to Sophie Borwein.
