Gender Gap in AI Skepticism: New Research Shows Women More Wary Than Men in Higher Education

Unveiling the Latest Research on Gender Differences in AI Perceptions

The artificial intelligence (AI) landscape is evolving rapidly, but new research highlights a significant gender gap in AI skepticism, with women expressing greater wariness than men. 9 69 Published on January 20, 2026, in PNAS Nexus, the study titled "Explaining Women's Skepticism toward Artificial Intelligence: The Role of Risk Orientation and Risk Exposure" draws from surveys of approximately 3,000 respondents in the United States and Canada. Researchers Beatrice Magistro from Northeastern University, R. Michael Alvarez from Caltech, and colleagues found that women are consistently more skeptical of AI, particularly regarding its workplace adoption, perceiving it as riskier especially under economic uncertainty. 67

This isn't mere opinion; women were 11% more likely than men to believe AI's risks outweigh its benefits. When asked about AI's societal impacts, women were 6-7 percentage points more likely to express uncertainty or deny benefits altogether. The gap vanishes, however, when researchers presented scenarios guaranteeing job gains from AI, suggesting uncertainty aversion is central. 66

Decoding Risk Aversion: A Core Driver of the Gender Gap

Risk aversion, a well-documented gender difference in behavioral economics, plays a pivotal role. In the study, participants faced a choice between a guaranteed $1,000 or a 50% chance of $2,000 (or nothing). Women overwhelmingly preferred the sure thing, reflecting higher general risk aversion. This trait extends to AI: women view technologies like generative AI (GenAI) or automation as potential threats to stability. 69

"Basically, when women are certain about the employment effects, the gender gap in support for AI disappears. So it really seems to be about aversion to uncertainty," Magistro explained. 66 Socialization contributes—girls are less encouraged toward STEM (science, technology, engineering, mathematics) fields, limiting exposure to AI's upsides and reinforcing caution.

AI Exposure and Job Vulnerability: Women on the Front Lines

Beyond innate aversion, women face disproportionate exposure to AI's disruptions. Occupations with high AI substitutability—like administrative roles, often female-dominated—risk automation. Conversely, AI-complementary jobs (e.g., AI oversight) skew male. Globally, women comprise only 25% of the tech workforce and under 20% of senior AI roles, per 2025 Women in Tech data. 69

In South Africa, where unemployment hovers around 32% (Stats SA, 2025), AI could exacerbate inequalities. Higher education sectors, employing many women in support and teaching roles, mirror this. For instance, administrative staff at universities like the University of Cape Town (UCT) or University of the Witwatersrand (Wits) may worry about AI tools replacing routine tasks.

Survey Methodology and Telling Statistics

Conducted via YouGov opt-in panels in November 2023, the survey measured AI attitudes on 0-10 scales, risk preferences via lotteries, and open-ended responses. Key stats:

• Women: 11% higher likelihood risks > benefits.
• Women: 6-7% more "don't know" on benefits.
• Gap closes to zero with guaranteed positives.
• Less risk-averse women match men's views.

These findings, funded partly by Caltech's Linde Center, underscore policy needs. 67

South African Higher Education: Contextualizing the Global Gap

In South Africa, AI integration in universities lags amid infrastructure gaps and inequality. A Council on Higher Education (CHE) Kagisano report (2025) notes top unis like Stellenbosch and UJ use AI more, but globally ranked institutions outpace others. Gender dynamics amplify skepticism: women, 55% of SA university students (DHET 2025), may hesitate adopting GenAI due to ethical concerns or bias fears. 44

A University of the Free State (UFS) study (2025) on private HE found minimal gender differences in AI attitudes among postgrads, but overall knowledge gaps persist. Tailored strategies, as suggested in UWC research (2025), could boost adoption. 62 UWC AI Adoption Study

Local Studies and Academic Perceptions in SA Universities

South African academics show cautious optimism toward GenAI for research efficiency and personalized teaching, per a 2025 study interviewing 16 across disciplines. 68 Sciences embrace it for data tasks; humanities worry about integrity. No explicit gender splits, but parallels global trends: women academics, often in humanities, voice stronger ethical qualms.

A systematic review (Jan 2026) on gender in student AI use across HE reveals mixed results—men adopt faster, women prioritize ethics. 43 At SA institutions like UKZN or NWU, this affects curriculum: prompt engineering courses could bridge gaps.

Implications for Careers in Higher Education

For aspiring academics, this skepticism influences paths. Women may shy from AI-heavy roles like data science lecturing, widening underrepresentation (only 30% female STEM profs in SA, NRF 2025). Explore research assistant jobs or crafting AI-savvy CVs to compete.

AI could displace adjuncts but create postdoc opportunities in AI ethics. Check postdoc positions blending AI and gender studies.

Challenges and Ethical Dilemmas in SA Contexts

Challenges include digital divides: rural unis like Fort Hare face bandwidth issues, hitting women students harder. Bias in AI—trained on Western data—risks cultural insensitivity in decolonizing curricula.

• Policy gaps: Few unis have AI guidelines.
• Inequality: GenAI aids English proficiency but homogenizes knowledge.
• Trust erosion: Hallucinations undermine reliability.

Solutions and Initiatives Bridging the Divide

Initiatives like Women in AI Southern Africa (@womeninai_sa on X) promote inclusion via events at UCT and Wits.Women in AI SA Policymakers advocate retraining, bias audits. Unis implement AI literacy: UJ's workshops demystify tools.

Steps for unis:

1. Gender-sensitive training.
2. Ethical frameworks.
3. Mentor programs.

Photo by Nico Smit on Unsplash

Future Outlook: Trends Shaping 2026 and Beyond

By 2026, expect SA HE policies mandating AI modules. Global trust may rise with transparent AI, narrowing gaps. Monitor CHE updates; opportunities in university jobs grow.

Optimism: Certain benefits could flip skepticism to enthusiasm, fostering equitable AI in education.