You have 1 article left to read this month before you need to register a free LeadDev.com account.
Estimated reading time: 5 minutes
Key takeaways:
- Women use AI at work 22% less than men, driven by less manager encouragement and training, not a skills gap.
- Female engineers who use AI get rated 13% less competent by peers, versus 6% for men.
- Left unchecked, this becomes a career gap. Fix it by measuring adoption directly rather than assuming a level playing field.
AI is changing the future of work – but old workplace inequalities are coming along for the ride.
The research paper Global Evidence on Gender Gaps and Generative AI Over Time argues that while generative AI can significantly boost productivity, it can also amplify existing inequalities. This means the technology may reinforce the very workplace disadvantages some say it aims to disrupt.
Women are adopting AI at work more slowly than men, according to data from Lean In. After surveying more than 1,000 Americans, the global nonprofit found that women are 22% less likely than men to use AI regularly at work.
Hilary DeCesare, founder and CEO of The reLaunch Company, says AI adoption is not happening on a level playing field when it comes to gender.
“The data backs up what a lot of us have felt in the room for a while now,” she adds.
That is that, as AI becomes a key driver of productivity, visibility, and career progression, the gap between who uses it and who doesn’t could determine who gets ahead, and who gets left behind.
Your inbox, upgraded.
Receive weekly engineering insights to level up your leadership approach.
New tech, same story
Women aren’t being given the same conditions to adopt AI as their male counterparts. Existing workplace gender biases are now showing up in how this new technology is used and rewarded.
The Global Evidence on Gender Gaps and Generative AI Over Time paper suggests women may continue to lag behind men in AI adoption because they receive less training and institutional support.
Lean In found that men are significantly more likely to say their managers encouraged them to use AI at work than women. Men are also 27% more likely to say their managers praised them for using the technology.
Deloitte’s research paints a similar picture: 93% of men say their company actively encourages AI use, compared to 84% of women, and 91% of men have received formal training on it, compared to 72% of women.
Jossie Haines, a former VP of software, believes that people are at very different stages of AI adoption, with varying levels of experience, confidence, and skepticism.
“Someone who’s dubious about AI adoption needs something completely different from someone already using it daily who wants to go further. Most enablement I see treats everyone the same, which means it lands for the people who were going to get there anyway. Working out where each person actually is, and what would move them up a rung, is an important management job,” Haines explains.
The AI competence penalty
Global Evidence on Gender Gaps and Generative AI Over Time also found that women may face greater scrutiny when they use AI at work.
This is known as the AI competence penalty, whereby women are more likely to be perceived as incompetent, creating a fear of using new technology.
This is explored in The hidden penalty of using AI at work research by Harvard Business Review. The study of more than 1,000 engineers found that women face a steeper penalty for using AI at work.
Engineers rated AI-assisted code 9% worse than identical code they believed engineers had written alone. When researchers revealed the coder’s gender, respondents rated female engineers who used AI 13% less competent, compared with 6% for male engineers.
The penalty was even greater among male engineers who did not use AI themselves, who rated female colleagues who used it 26% less competent.
“Run that gap forward and you get a loop,” says DeCesare. “A man tries AI, gets encouragement going in, gets praised coming out, and that praise becomes evidence. His brain registers, ‘this is who I am now, someone who uses this well.’ A woman tries the same tool, gets less encouragement and less credit, so the evidence never fully lands.”
“When a new tool shows up, the calculation isn’t ‘will this help me?’ it’s ‘what does it cost me if this goes wrong in front of the wrong person?’ she adds.
When AI fluency becomes a proxy for speed, judgment, and leadership potential, a gender gap in AI use can become a career gap, DeCesare says.
Women who use AI less may miss out on high-profile projects, recognition, and opportunities that lead to promotions.
More like this
Fixing the gender gap
AI adoption isn’t just about who knows how to use ChatGPT or Copilot. It’s about who feels permitted to experiment, who gets rewarded for doing so, and who risks their professional reputation by getting it wrong.
Haines argues that organizations should stop simply pushing AI adoption and instead focus on removing the career penalty associated with using AI.
“Start by measuring the exact gap instead of assuming your culture is the exception,” Decrease says.
New York • September 8 & 9, 2027
Loved LDX3 New York? Pre-sale tickets for 2027 are now available.
Beyond that:
- Look past who technically has access to AI tools and examine which managers are actively encouraging and supporting people to use them. Access without explicit encouragement does not create equity.
- Don’t tie AI use to performance evaluations, as this can discourage experimentation and learning – particularly among people who already face competence penalties for visible AI use.
- Focus on enabling AI skeptics, chiefly senior engineers and managers, because non-adopters may be more likely to judge colleagues harshly for using AI.
- Fix the praise and credit gap deliberately. If men are publicly recognized for using AI while women doing equivalent work receive no acknowledgement, managers need to correct that behavior openly and in real time.
- Recognize AI enablement work as career impact. Creating guardrails, review standards, and processes can be more valuable than simply producing the most AI-generated output, but this work needs to be visible in performance and promotion decisions.
- Measure adoption more carefully, breaking down AI fluency by gender and seniority rather than relying on a single organization-wide adoption figure that can hide inequalities.
