Women in Tech Discrimination Rises to 47% as AI Gap Widens

Discrimination reports go up

The technology industry has spent ten years talking about getting more women into the field. Data published this autumn shows that much of that talk has not changed the experience of women already working in it. ISACA's 2026 Tech Workplace and Culture survey, released on 13 October, found that 47% of women in tech say they have faced gender discrimination at work. In 2024 the figure was 42%.25 ISACA read the five-point rise as a sign that growing awareness of the problem has not produced change.2

The survey reached 1,025 tech workers around the world between 17 July and 3 August 2026.2 On almost every measure, women reported a worse working life than men. Fifty-two percent of women were satisfied with workplace culture, compared with 65% of men. For job security the figures were 59% and 69%, and for career progression 67% and 73%.36 The widest gap was over influence. Only 53% of women felt they had enough authority to shape company strategy, against 66% of men.2

Burnout follows the same pattern. Across the whole sample, 59% of workers reported work-related stress or burnout. Among women the figure was 64%, compared with 57% of men.2 Unclear career progression was the most commonly cited obstacle to advancement. Forty-one percent of women named it, compared with 33% of men.9 Coverage of the survey treated it as a single picture with no dispute over its numbers, and several outlets reprinted the same figures nearly word for word.369

The survey also found optimism that is hard to square with the rest of the results. Seventy-one percent of both men and women expect more women to enter tech over the next five years.3 ISACA's Julia Hermann called the results mixed and said employers' actions, especially hiring programmes aimed at women, have not kept up with awareness.2

AI hiring is creating a new gap

The ISACA results would be worrying in a stable industry, but tech is in the middle of an AI hiring boom. That boom matters most for the long-term picture, and it is where the coverage agrees most strongly.

LinkedIn data reported by the Guardian shows women made up about a quarter of new hires into AI roles over the past year. Women accounted for half of hires into non-AI roles over the same period, and only 13% of AI hires at executive level.8 This matters because AI pays more. LinkedIn found that AI jobs pay more than twice as much as roles that do not involve AI on average, and that median pay for men across all AI occupations is $45,000 higher than for women. Part of that gap comes from women being clustered in lower-paid work such as data annotation.14

The World Economic Forum's 2026 Global Gender Gap report reached a similar conclusion: women make up nearly half of data annotators but only about a fifth of AI developers and deployment engineers.4 UN Women's Gender Snapshot 2026, released at the UN General Assembly, estimates that women are about 22% of AI professionals worldwide. It also says they hold fewer than 14% of senior leadership roles and are around 20% less likely than men to use generative AI tools.10

The exact share of women in AI depends on who is counting. One compilation of statistics found that women are 29.4% of people listing AI engineering skills and 39% of those in data and AI roles more broadly.1 The International Labour Organization put women at about 30% of the global AI workforce in 2022, only four points higher than in 2016.15 These figures conflict less than they seem to, because each one counts a different group. They all point the same way, though. Women are well represented at the bottom of the AI labour market and become scarcer higher up. That is why the most persuasive comparison here is between AI hires and other hires, and not any single headcount.

Exposed at both ends

Several recent studies describe a two-sided problem. Women are missing from the jobs that build AI and over-represented in the jobs AI is most likely to replace. The ILO has found that women's concentration in clerical and administrative work makes them more exposed to generative AI than men.15 Research from Brookings and the Centre for the Governance of AI identified about 6 million US workers who would find it hardest to recover from losing their jobs to AI. Eighty-six percent of them are women, mostly in clerical and administrative roles.17 One Brookings researcher said this reflects the jobs women hold, not their ability.17

The WEF reports that women make up 57% of US workers in roles likely to be disrupted by generative AI.12 McKinsey, as cited by Forbes, estimates that women are 1.6 times as likely as men to be among the roughly 770,000 US workers who may need to change occupation each year because of AI and automation. Forbes also pointed to a quieter warning sign. At entry level, managers encourage 33% of men to use AI tools but only 21% of women.

It is less clear whether women are falling behind as users of AI. A Harvard Business School analysis pooling 54 sources found that 47.8% of men use AI at work, compared with 39.3% of women.1 A Boston Consulting Group study of tech workers specifically found that women use generative AI at work slightly more often than men, at 68% versus 66%.11 The likeliest explanation is that women already inside technical jobs are adopting AI at least as quickly as men. The wider gap reflects the jobs women hold across the whole economy. If that is right, the usual advice to tell women to learn AI tools misses the main issue, which is which jobs they can get.

Gains reversing

The most worrying evidence is that past progress is being undone. McKinsey research cited by the WEF found that women's share of tech roles in Europe fell from 22% in 2023 to 19% in 2026, partly because layoffs hit roles where women were concentrated.12 Findem's earlier analysis found that women were about a third of employees at FAANG companies but 44% of those laid off. Many of the cuts fell on HR, recruiting and marketing teams, where women had made gains.13

The old remedies are also weaker than before. The Guardian reported that companies have cut back sharply on diversity programmes in response to the political climate. As a result, fewer employers have formal systems for hiring and promoting women, and advocates inside companies say women-focused initiatives are harder to keep going.14 Brenda Darden Wilkerson of AnitaB.org called the trend backsliding. She said fast AI hiring still runs through the same networks and referrals that have long favoured men.14 One statistics roundup noted that company-level data has become harder to find since Google, Microsoft and Meta stopped publishing diversity reports.1

Women inside AI teams say the work itself creates barriers. Jayeeta Putatunda, an AI engineering lead, told the Guardian that after four months of maternity leave she came back to find completely different frameworks and models. She said women leave AI because they lack support, not because they lack ability.14

The pipeline problem

The barriers start before women are hired and continue once they are. Women earned roughly 23% to 25% of US computing bachelor's degrees in recent Computing Research Association surveys.1 LeanIn.Org and McKinsey report that 93 women are promoted to manager for every 100 men, and only 60 Black women. Women's share of C-suite roles has stayed at 29%.1 Location adds to the gap. In SmartAsset's 2026 city rankings, San Antonio came 42nd of the 50 largest US cities: women make up 23% of its tech workforce and earn about 84.6% of what men in tech earn there.7

When asked what would help most, ISACA's respondents named early intervention. Sixty-nine percent of women want more female mentors and role models.9 Organisations working on the pipeline are also changing leadership. Susan Gouijnstook became CEO of Girls Who Code in September.18

What it means

Taken together, the data points in one direction. The tech industry's gender gap is not slowly closing. It is moving from older parts of the industry into AI, where pay, power and influence over how the technology is designed are now concentrated. Mentoring and role models, the fixes respondents asked for, are worthwhile but come late in a process where the main problems are in hiring, promotion and layoffs. The rise in reported discrimination at a time of supposed awareness suggests that awareness is no longer the main obstacle. An advocate quoted by the Guardian warned that a gap in participation can turn into a gap in power.14 Based on this autumn's data, that is already starting to happen.