Researchers from Open University and Politecnico di Torino find female-dominated occupations face higher exposure to Large Language Models
Bot Mutiny |
A gendered analysis of labor market data shows that while AI exposure for men is concentrated in high-wage roles, women in low-skilled and low-paid positions face significant exposure to task automation and job restructuring.
Gender inequality is a structural reality of the labor market. It dictates lifetime earnings and economic security. New research suggests that the deployment of artificial intelligence is not hitting the workforce evenly. A study published on arxiv.org on September 18, 2026, details how AI exposure varies across male- and female-dominated occupations. The paper, authored by researchers from the Centre for Protecting Women Online at the Open University, Politecnico di Torino, Nokia Bell Labs, and the Fawcett Society, is a preprint that has not yet been peer-reviewed. The research team linked O*NET occupational data with U.S. Census gender-disaggregated income distributions. They used two specific indices to measure risk: the Anthropic Index for Large Language Model (LLM) exposure and the Artificial Intelligence Index for broader innovation like robotics. ## The Wage Gap in Exposure
The data shows a distinct split in how tech hits different genders. In male-dominated fields, AI exposure is concentrated in higher-skilled and higher-paid roles. For men, the tech often acts as a complement to high-skilled work. Female-dominated occupations don't have that buffer. The study found that exposure in these fields is uniform across the entire spectrum. It doesn't matter if the job is high-paid or low-paid. Women in low-skilled, low-wage positions are just as exposed to AI-driven changes as those at the top. ## LLMs vs. Robotics
The type of AI matters. The researchers found that LLM-related exposure is higher in female-dominated occupations. Conversely, exposure to broader AI innovation and robotics remains concentrated in male-dominated sectors. The authors state that exposure doesn't always mean immediate firing. It captures the extent to which tasks are likely to be affected. However, the triangulation of this data with existing literature paints a grim picture for vulnerable workers. Women in lower-skilled positions have weaker bargaining power and limited access to training. For them, AI exposure is less likely to result in "productivity-enhancing augmentation." It's more likely to mean the reorganization or substitution of their tasks. This leads to reduced wages and stalled career progression. ## The Adoption Gap
The study cites previous data showing that men already use generative AI tools more than women, 50% versus 37%. This gap is driven by differences in technical knowledge, trust, and privacy concerns. If women participate less in AI use, these systems are trained on data that doesn't reflect their needs. It creates a cycle where the technology reinforces existing economic disparities. Lower-skilled workers already participate in adult learning at lower rates. Without intervention, the researchers suggest AI adoption will generate uneven effects. It'll leave women in vulnerable positions facing the highest risks of displacement and wage suppression.
References
(2026). When AI Enters the Workplace, Who Faces Greater Risks? A Gendered Analysis. arxiv.org. https://doi.org/10.54394/00033798