Research briefing

The AI confidence gap for women

Women’s lower adoption of generative AI is not evidence of lower ability. The research points to differences in encouragement, risk and the conditions for experimentation.

Across studies covering more than 100 countries, about 39% of women had adopted generative AI compared with 48% of men. The useful response is not more hype; it is safer, practical opportunities to learn by doing.

39% vs 48%women’s and men’s reported adoption
32% more likelyfor women to worry AI use looks like cheating
23% more likelyfor men to be encouraged by a manager
8× growthin the small-business adoption gap since 2019

What the evidence says

A Harvard Business School analysis synthesizing research from more than 100 countries found a persistent gender difference in generative AI adoption. The gap appeared across contexts rather than being explained by one country, industry or product.

Lean In’s workplace research adds an important mechanism: women report less encouragement to use AI and greater concern that using it will be interpreted as cheating. That changes the risk calculation. Trying a new tool is easier when experimentation is rewarded and harder when it may be judged as taking a shortcut.

JPMorganChase Institute research on small business owners found that the gender gap in adoption had grown substantially since 2019. That matters because AI capability is becoming part of ordinary business operations, not a specialist technical function.

Why confidence is not a personality trait

Confidence is often treated as something an individual either possesses or lacks. In practice, it is shaped by conditions: whether mistakes are safe, whether help is available, whether peers are experimenting too and whether the work feels relevant.

A hands-on workshop changes those conditions. Participants start together, practise on a shared example and then use the tool on work they already understand. That makes it easier to judge the output and easier to see a reason to use the skill again.

What helps close the gap

  • Training built around real work rather than generic demonstrations
  • Clear safety boundaries and permission to ask basic questions
  • Peer examples that expand the set of visible use cases
  • Small, repeatable tasks that create evidence of personal capability
  • Follow-up that turns a workshop experience into a working habit

Sources

  1. Harvard Business School: global evidence on the gender gap in generative AI
  2. Lean In: workplace encouragement, praise and concern about “cheating”
  3. JPMorganChase Institute: AI adoption among small business owners