Women are adopting artificial intelligence at a slower rate than men despite being disproportionately represented in jobs most exposed to AI disruption, raising concerns that the technology could deepen existing gender inequalities in the workplace and limit the economic gains expected from the AI boom.
A Harvard Business School paper found that the AI adoption gap between men and women was 22%, based on adoption rates of 47.8% among men and 39.4% among women.
The gap has since narrowed to 16%, according to the analysis by Tamryn Geldenhuys and Tamsin Mackay, but remains across income, education and occupational groups. The authors argue that the disparity is less about women’s ability to use the technology than unequal access and perceptions around AI.
“The pattern holds across income levels, education and occupation, which means that really the gap isn’t necessarily caused by ability but rather by access and perception,” the authors say.
The problem is that the roles most likely to be affected by AI are the ones most often held by women.
The economic stakes are growing as AI becomes embedded in businesses and workplaces.
AI is projected to generate nearly $20 trillion in global economic gains by 2030, but those gains could be constrained if women remain underrepresented in the technology’s development and adoption.
Women account for 57% of roles likely to be disrupted by AI, compared with 43% for men, according to research by the World Economic Forum in partnership with LinkedIn.
At the same time, women represent only 29% of AI-skilled workers, compared with 71% of men, according to a Randstad survey cited by the authors.
That creates a widening exposure gap: women are heavily represented in occupations being reshaped by AI but are less represented among workers equipped with the skills to use the technology effectively.
The barriers are not limited to access to tools. The authors point to perceptions that using AI can amount to “cheating” and that the technology is not necessarily a valuable workplace collaborator.
Women can also face greater professional penalties when their use of AI becomes known.
A Harvard Business Review study found that engineers believed to have used AI were rated about 9% less competent for identical work. The penalty was higher for women, at 14%, compared with 6% for men.
The disparity extends into the AI industry itself.
A Women at the Table analysis prepared for the UN Commission on Science and Technology for Development found that women hold only 31% of jobs at major AI companies and about one-third of AI research positions.
The analysis also links the underrepresentation of women in datasets and design processes to measurable failures in AI performance.
AI adoption strategist Mara Bolis has warned of the risk of creating a “two-tiered AI economy” if women are not more actively included in the technology’s development and use.
The authors argue that organisations should stop presenting AI primarily as a technical skill and instead teach it as a communication capability.
Effective use of generative AI often depends on how clearly users frame questions, provide context, give instructions and evaluate the resulting output.
Simply giving employees access to AI tools may not be enough. The authors argue that organisations need deliberate, inclusive training that gives women practical experience and clear frameworks for developing and checking prompts.
“Those that treat AI training as an inclusive and ongoing investment will be able to keep and advance their talent instead of watching old inequities being rebuilt into the way work is done next,” the authors assert.
The slower adoption rate could also present an opportunity. Greater caution around AI may reflect stronger awareness of governance, ethics and risk, qualities that could become increasingly important as businesses integrate AI into core operations.
For companies, the economic issue is increasingly tied to talent.
Organisations that invest in inclusive and ongoing AI training could help more women adapt to changing jobs while retaining experienced employees and preventing existing workplace inequalities from being reproduced in the next generation of work.
