CLOSING THE GENDER GAP IN AI ADOPTION. WHY DO WOMEN ADOPT AT LOWER RATES THAN MEN?

Women are 22% less likely to use AI than men. What's more, they're more likely to consider using AI as cheating. (1).

So while AI has the potential to increase US economic output by 20% over the next decade (Baily, Brynjolfsson, and Korinek, 2023), given that women represent about half the US workforce, a large gap in adoption could result in hundreds-of-billions of dollars of lost productivity. (2)

Moreover, the less women use AI, the more we perpetuate the gender biases that already exist in AI models. (3) (Want to see it firsthand? Ask AI for specific advice for a woman, and then the very same advice for a man).

Research indicates that a lack of access does not cause the adoption gap. In fact, when women do have equal access, adoption still trails men's. (4)

Why do we see this gender imbalance in AI adoption? According to a meta-study published in this HBR working paper cited earlier, there are 4 reasons -- differences in:

o   Lower familiarity and knowledge of AI Tools

o   Lower confidence and persistence of use (Self-Efficacy). Women are less likely to have confidence and, therefore, are less likely to persist with prompts.

o   Knowledge, familiarity, and skill

o   Beliefs that using AI is "unethical" or "cheating."

The gender gap in AI adoption is greatest with younger women, especially those outside tech. In a 2024 BCG convenience sample of 6558 tech employees, senior women in tech adopt AI about the same or slightly higher than senior men. Less experienced women in tech and women outside of tech adopt AI at lower rates than their male counterparts.

Why do senior women in tech adopt at higher rates? The BCG study suggests that they perceive it as more critical to their future job success, have a higher tolerance for using AI without an explicit company policy, and have greater confidence in their AI skills.

The authors of this study suggest that women have had greater pressure to prove their competency in an industry still disproportionately male. One respondent summed it up this way: "Senior women in tech have broken barriers to get where they are, but they still feel they need to prove themselves and take more initiative than men to be abreast with what's important for their careers [such as GenAI]."

So, how do we close this gap in AI adoption?

The obvious answer might be to reduce systemic bias. But why boil the ocean? Instead, consider behavioral design, where you can make immediate, tangible progress by implementing smaller, easier, high-impact behavioral interventions you can measure.

• 
  

  -- One of the best predictors of high AI adoption is how much you see others similar to you adopting it. They're not exactly peers but near-peers, people slightly more technically competent who are open to innovation and have higher social capital than their peers. I wrote about the importance of early adopters and near-peers in an earlier post. Showcase the successes of female early adopters equally with men, even if they don't represent 50% of your early adopters. Raise the visibility of their work through lunch and learns, Q&A's, whitepapers, and demo days. Announce successes through internal digital communications and publicly on LinkedIn.

• Hire Female AI Trainers --  In a recent study published by the Journal of Marketing Research, women of equal aptitude as men underperformed in quantitative classes. However, their performance significantly improved when female instructors taught them. Women's confidence and interest in stereotypical male fields improved, too. (6) Doesn't this disadvantage men? Interestingly, the instructor's gender did not impact male students' performance. (6)

  
  

• Focus on Small, Quick Wins. We're more likely to start any new task or undertake any new routine if we think we will be successful. Lower-stakes work builds mastery while reducing anxiety. The bigger the initiative, the greater the anxiety, and the more resistance. Start small. Encourage the use of AI for summarizing a meeting or research, drafting memos, or conducting research on competitors.

• Redesign Your AI Policy.

  1. Frame AI usage more positively. Many policies highlight the sanctions against the inappropriate use of AI but don't define appropriate use. I've written about how this results in shadow AI use. Research suggests that where AI is discouraged or unclear, younger women are less likely than men to use it (BCG Study)

  2. Frame AI as a thought partner (not a shortcut or a replacement). Women are more likely to feel that they are cheating because they have typically had to prove their competency in the workplace (BCG study). We often prove competency through effort. Removing effort can delegitimize women's sense of their own competency. Instead, highlight the role of the unique expertise required to make AI a force multiplier -- judgment, creativity, and domain expertise. Let them know that frequency of use builds competency.

  
  

• Measure It.  Finally, anonymize your data on AI usage and measure adoption by men and women. Daily usage is the behavior that best captures adoption for most job functions. Consider conducting periodic surveys to measure how confident your employees are with using AI and compare the change in confidence between men and women.

Don't risk losing substantial productivity gains because half your employees use AI less often and less effectively. Close the gender gap. Start today. These suggestions will create a culture in which women feel just as empowered, competent, and invited to use AI as men and deliver the productivity gains your board expects.

Sources

(1) Global Evidence on Gender Gaps and Generative AI, Harvard Business Review Working Paper 2024, 2025 by Nicholas G. Otis, Solène Delecourt, Katelyn Cranney, Rembrand Koning. This report synthesized data from 18 studies covering more than 140,000 users of popular generative AI platforms

(2) How Generative AI Will Power the Coming Productivity Boom.", Brookings Institution 2023.

(3) "Sampling bias in entrepreneurial experiments." Management Science 2023

(4) Global Evidence on Gender Gaps and Generative AI, Harvard Business Review Working Paper 2024, 2025 by Nicholas G. Otis, Solène Delecourt, Katelyn Cranney, Rembrand Koning.

(5) Global Evidence on Gender Gaps and Generative AI, Harvard Business Review Working Paper 2024, 2025 by Nicholas G. Otis, Solène Delecourt, Katelyn Cranney, REmbrand Konin