Many women lawyers are deciding whether to set aside scarce hours this quarter to learn AI tools properly, and what to ask their employer for first. We wanted to answer that from the evidence, so we read the research on women, work and generative AI, including the studies that complicate the picture.
Our view is that the gap in AI use between women and men comes mostly from conditions at work: permission, training, encouragement and credit. A team can change each of these, and a woman lawyer can ask for each of them. The evidence on lawyers specifically is thin, and we say where.
63%
Solicitors who are women
People in SRA-regulated law firms in England and Wales, summer 2025. Solicitors Regulation Authority [1]
39.3%
Women using generative AI
Against 47.8% of men, across 76 studies and surveys. Cranney, Delecourt and Koning [4]
16 points
Gap within one survey
How much less likely women were than men to have used ChatGPT for work, among 18,000 Danish workers in 11 occupations. Humlum and Vestergaard [5][8]
Who is the profession now?
In each jurisdiction below, women are the majority of solicitors or of new law students.
| Jurisdiction | Figure | Group covered | Source |
|---|---|---|---|
| England and Wales | 55% of lawyers are women, up from 48% in 2015; 63% of solicitors are women | People in SRA-regulated law firms, data collected summer 2025 | [1] |
| England and Wales | 35% of full-equity partners are women | People in SRA-regulated law firms, summer 2025 | [1] |
| Australia | 56% of practising solicitors are women; 64% of solicitors admitted for five years or less are women | Practising solicitors, October 2024 | [2] |
| United States | Women have been the majority of first-year law students every year since 2015 | First-year law school classes | [3] |
Women are most solicitors but a minority of the most senior partners. In England and Wales law firms, women are 63% of solicitors but 35% of full-equity partners [1]. If skill with AI comes to affect who is given senior and visible work, it could widen or narrow the distance between those two figures, depending on who learns the skill and who is credited for it.
Exhibit 1
Women are most solicitors, and about a third of full-equity partners
Women as a share of each group, per cent. England and Wales: people in SRA-regulated law firms, summer 2025. Australia: practising solicitors, October 2024. The dashed line marks 50%.
Sources: Solicitors Regulation Authority [1]; Urbis for the Law Society of New South Wales [2].
How large is the AI gap?
Across many occupations, fewer women than men report using generative AI. The largest body of evidence is a review by Katelyn Cranney, Solène Delecourt and Rembrand Koning, a Harvard Business School working paper updated in May 2026. It combines 76 sources, including academic studies, industry surveys and government reports, covering more than 300,000 people in over 100 countries. Giving larger studies more weight, 47.8% of men and 39.3% of women reported using generative AI. The gap has narrowed since ChatGPT launched, but since early 2025 it appears to have stopped narrowing [4].
The gap between women's and men's use also appears within a single occupation [4]. A survey of 18,000 Danish workers in 11 occupations where AI tools can do part of the work, including legal professionals, found that women were 16 percentage points less likely than men to have used ChatGPT for work [5][8].
Hiring shows a gap too: women made up 26% of US hires into AI jobs in 2025, against 50% of hires into other jobs [9]. LinkedIn counts AI jobs by occupation and AI engineering skills, so the figure is about who builds AI systems more than who uses AI in legal practice.
We found no reliable published figure comparing women lawyers' AI use with men's. In a 2024 survey of 90 women in law published by Linklaters, 18% named fear and lack of confidence as holding women back, while many said they had seen no slower take-up among female colleagues [10]. So the gap is well measured across occupations and barely measured in law.
Why do women use AI less?
The research points mostly to conditions at work. Cranney and colleagues sort the causes into five groups of barriers, shown below with related survey evidence. The Lean In figures come from a 2026 survey of 1,015 adults in the United States across all occupations [4][6].
| Barrier | What the review found | Related evidence |
|---|---|---|
| Knowledge | Women report knowing less about AI and score lower on objective tests | 17% of the Linklaters respondents said they had a knowledge gap [10] |
| Doubt that it helps | Women are less sure AI will help their careers or productivity | No legal-specific figure found |
| Support and training | Women report less employer encouragement and less training | 30% of women and 37% of men said their manager encourages them to use AI [6]; 54% of legal professionals in one survey said their firm gives no training on responsible AI use and plans none [11] |
| How it looks to others | Women worry more that AI use will look like cheating or low ability | 29% of women and 22% of men worried about being seen as cheating [6] |
| Trust and risk | Women report stronger concerns about data security and risk | 37% of the Linklaters respondents feared AI's reliability [10] |
Clear permission may change how much women intend to use AI. At NHH Norwegian School of Economics, female students used AI tools less than male students. When the 595 students imagined that AI use was explicitly allowed in class, the gap in intended use disappeared, with over 80 percent of both men and women intending to use the tools. When they imagined that AI was explicitly banned, the gap in intended use widened again, most among top-performing women [7]. The Danish workers likewise reported that employer restrictions and a perceived need for training often held them back [5].
The worry about being judged has some basis. In research on software engineers, identical AI-assisted work received lower competence ratings when the engineer was female, and female engineers anticipated this [4]. In the Lean In survey, among people who had used AI at work, 23% of men and 18% of women had been praised for it [6]. The question is who gets credit for using AI as well as who uses it.
Exhibit 2
Women report less encouragement and less praise for using AI at work, and more worry about how it looks
Share of US adults across all occupations, per cent, 2026. Praise is among people who had used AI at work.
Survey of 1,015 adults in the United States. Source: Lean In [6].
Lack of time also limits learning. Linklaters respondents said caring responsibilities, maternity leave and extra career-building activities cut into time to learn, and that practical experience (24%), training (20%) and mentoring (20%) would help women adopt AI [10].
Most of these barriers will not go away by themselves. The review notes that the barriers of knowledge and doubt about usefulness have shrunk as AI tools have spread. The other three depend on how workplaces are organised, what colleagues expect, and trust, and are "less likely to fade on their own" [4]. Waiting for the gap to close by itself is therefore unlikely to work.
What does it mean for careers?
AI use is becoming ordinary in legal work. In 8am's 2026 Legal Industry Report, based on a survey of more than 1,300 legal professionals in the United States, Canada and Puerto Rico, mostly in solo practices and small firms, 69% used general-purpose generative AI tools for work, up from 31% a year earlier [11]. Once most lawyers use the tools, using them sets no one apart. What sets a lawyer apart is how well she uses them, and whether her employer knows it.
Managers say they value AI skill, especially in women. In the NHH study, 1,134 managers rated top female students with AI expertise 7.6 percent higher than those without, while male candidates did not receive the same higher rating [7]. This was a hypothetical rating exercise and does not show who the managers hired.
Women may receive less credit for the same use. The Lean In praise figures and the software engineer study both suggest that women who use AI can receive less recognition for it than men [4][6]. Neither study is of lawyers, so neither shows what happens in a legal team.
Learning the tools may therefore not be enough on its own: a woman lawyer's use also needs to be permitted, visible and attributed to her. Lawyers know this from any piece of drafting: the work counts in a review when the supervising partner knows who did it.
A worked example
Mei Lin is senior counsel at Ostrava Freight, a fictional logistics company with a legal team of six. The team receives about 40 requests a month for non-disclosure agreements, and each waits an average of four working days for a first response. Mei emails her general counsel asking for two hours a week for eight weeks to build a triage checklist with the company's approved AI tool, using only invented agreements. The general counsel agrees.
In the first week Mei writes down who receives each request, who reviews it, and where it waits. Most of the delay is requests sitting unopened in a shared inbox, and two thirds are on the company's own template, unchanged. She asks the AI tool to draft a checklist that sorts requests into three groups (template unchanged, template edited, counterparty paper), and tests it against 30 invented agreements, correcting it until it sorts all 30 correctly. A paralegal now uses the checklist to approve unchanged template requests on the day they arrive. After eight weeks Mei sends her general counsel a one-page note with before and after turnaround times, and the note is cited in her annual review.
The email settles permission, the agreed hours protect her time, the invented data keeps client information out, and the note gets her credit.
Where the skeptics are right
Some of women's lower use may also be sound caution. Lawyers owe duties of confidentiality and competence, and AI tools can produce errors that look correct. A lawyer who holds back until she can check the output and protect client data is doing her job. The review recommends careful use supported by the employer, with clear policies, training and privacy safeguards [4]. The actions below follow that approach.
The evidence also has limits we want to state plainly. Most of these figures come from surveys, which show associations: women report less encouragement and less use, but the surveys do not prove that one causes the other. The review draws on 76 sources of different kinds, and the only legal-specific evidence on women and AI here is a 90-person survey from 2024 [4][10]. The NHH results come from students and from managers answering hypothetical questions [7]. The profession figures cover England and Wales, Australia and the United States only. The findings point the same way across sources, and the actions below cost little even if the gap is smaller in law.
What we'd ask
If you lead a legal team or a firm:
Most of the causes the review identified sit with the employer, so the larger changes are for teams [4].
- Publish a short AI policy that says what is allowed, with which tools and which data. In the NHH scenario, explicit permission closed the gap in intended use, and a ban widened it [7].
- Train everyone, with time set aside for practice. In a UK pilot cited by the review, short hands-on training substantially increased women's use [4].
- Ask managers to encourage AI use directly and equally. Men were more likely than women to report this encouragement [6].
- Give credit for careful use in reviews and work allocation. Linklaters recommends making AI literacy a factor in promotion decisions [10].
- Include the people who do the work when choosing tools and redesigning processes. In England and Wales and in Australia, most solicitors are women, so a tool chosen without women's input is chosen without most of its users [1][2].
If you are a woman lawyer deciding whether to start:
- Ask for permission in writing. A short email to your manager or general counsel naming the tool, the task and the data you will use settles whether you are allowed. Of the barriers, permission is the one this research tests most directly [5][7]. If you set the rules yourself, as a partner or sole practitioner, write them down for your team.
- Agree the time in advance. A fixed two hours a week is easier to protect than a plan to learn when work is quieter, and lack of time was a constraint Linklaters respondents raised [10].
- Practise on fictional data. Invented facts let you test a tool with no client information involved, which deals with the concern about data security [4].
- Pick one recurring problem in your team's work, and record how long it takes before you change anything.
- Share a written record of what you changed and what it saved, so your manager has something specific to credit.
Why we built the course
The research says the gap in AI use depends on permission, training, time and credit at work, and that it is unlikely to close on its own. We built AI Women in Law to give women lawyers training, set time to practise, and a piece of finished work to show for it. Over five weeks, participants redesign one piece of their team's work and build a first version of a tool for it. See the course.
References
- Solicitors Regulation Authority, Diversity in law firms' workforce, 11 March 2026.
- Urbis for the Law Society of New South Wales, 2024 National Profile of Solicitors, 13 June 2025.
- Law School Admission Council, LSAC's Knowledge Report: The Composition of the First-Year Law School Class and Enrollment Trends, 2025.
- HBS AI Institute, The Gender Divide in Generative AI: A Global Challenge [Update], 6 August 2026, summarising K. Cranney, S. Delecourt and R. Koning, "Global Evidence on Gender Gaps and Generative AI Over Time", Harvard Business School Working Paper No. 25-023, May 2026.
- A. Humlum and E. Vestergaard, The unequal adoption of ChatGPT exacerbates existing inequalities among workers, Proceedings of the National Academy of Sciences 122(1), January 2025 (abstract on the University of Copenhagen research portal).
- Lean In (M. Noble-Tolla), New research: Women use AI less often at work and get less credit, updated 5 June 2026.
- NHH Norwegian School of Economics, Reveals gender gap in AI tool usage among students (on D. Carvajal, C. Franco and S. Isaksson, "Will Artificial Intelligence Get in the Way of Achieving Gender Equality?"), 18 September 2024.
- A. Humlum and E. Vestergaard, The Adoption of ChatGPT, Becker Friedman Institute Working Paper No. 2024-50, April 2024 (revised April 2025).
- LinkedIn, New LinkedIn Research Finds Women Account for Just 26% of AI Hires as AI Jobs Surge, 18 August 2026.
- Linklaters LLP, No Woman Left Behind: Closing the AI Gender Gap in Law, 2024 (survey May to June 2024).
- R. Ambrogi, LawNext, AI Adoption Among Legal Professionals Has More Than Doubled in a Year, New 8am Report Finds, But Firms Lag Far Behind Individual Practitioners (reporting 8am, 2026 Legal Industry Report), 5 March 2026.