Everyone Got the Free Chatbot. Not Everyone Got the Rest
Free chatbot use is now near-equal for women and men. Paid and agentic tools are not. What the AI tool access gap means for women running businesses.

Women now use free chatbots almost as much as men. The gap moved up a tier, to the paid and agent tools that do heavier work, and those usually get handed out by someone else. If you own the business, you are the one handing them out, so here is a five-minute check and a one-month test.
The usage gap really did narrow, and that deserves saying first
Pew Research Center’s October 16 analysis, drawn from 5,119 US adults surveyed in February, puts chatbot use at 50% of men and 47% of women. ChatGPT specifically came in at 44% for both groups, identical. Two years earlier the split was 39% against 28%, on question wording Pew notes has since changed.
Women did not sit this one out. The free tier is most of the reason the entry barrier fell so fast.
Where the AI tool access gap actually sits now
Read further into the same Pew tables and the convergence stops. Gemini: 29% of men against 20% of women. Copilot: 22% against 13%. Claude: 9% against 4%. The single tool with no gap at all is the one with the most generous free tier. The gaps open on tools whose fuller versions often arrive through an employer, a paid seat, or a developer account.
The mechanism came up at the October 14 launch of When AI Met Gender: Gendered Impact of AI on the Labour Market, a King’s College London report by Dr Hyojin Seo and Professor Heejung Chung of the King’s Global Institute for Women’s Leadership. There, OECD labor market economist Anna Milanez described cases where premium and agentic tools went mainly to managers, who are more often men, and where generative tools reached technical teams and early volunteers first, who are also more often men. Part-time workers, more often women, had the least time at work to experiment. None of that is about willingness, and none of it is a confidence problem. It is allocation, which sits next to the question of who gets taught to use these tools at all.
Allocation is a decision somebody else makes.
The same shape at the top of the field
UNESCO’s outlook study on artificial intelligence and gender in South Asia, launched in New Delhi in July, found women listed as an author on 71.76% of AI publications between 2015 and 2025, yet holding only 26% of corresponding or primary authorship, the roles that carry research leadership. The same study places women largely in data preparation rather than model development, where the technical authority sits. Dr Bhavani Rao, of the UNESCO Chair on Gender Equality and Women’s Empowerment, described systems that have not been designed with them in mind
. Presence near the work, distance from the decision.
For an owner, nobody is allocating her a seat
Here the structure flips, and it cuts both ways. A woman who owns the business has no manager rationing licenses to her. She decides. That is genuine control, and the part of this gap she can close without waiting for anyone.
The constraint that replaces it is cash. Microsoft 365 Copilot Business, the add-on for business plans, lists at $21 per user per month (discounted to $18 as of November 2026), the enterprise version at $30, and the Agent 365 governance add-on is a further $15 on top of a qualifying base license, according to Microsoft’s own pricing pages. For a business of three that is a line item somebody has to justify, on top of every other AI subscription already running.
The firm-level data says that margin is thinner to begin with. JPMorganChase Institute, working from de-identified Chase business banking transactions, found 19.7% of male-owned firms had adopted AI by the end of 2025 against 17.2% of female-owned firms, and noted that female-owned firms are less often employer firms at all, 17% against 23.5%. Smaller firm, tighter capital, fewer people to spread a seat across.
Reluctance belongs here too, and not as a mood. The 2026 Omni Calculator study of 808 US workers and 403 executives, reported by Forbes on October 20, found 33% of executives expect AI tools to narrow the gender pay gap, against 11% of employees. Reverse the question and 31% of employees expect those tools to widen pay disparities, against 9% of executives. The people closest to the tools are least convinced they will be pointed in their favor, which is worth taking seriously rather than talking anyone out of.
A five-minute check: who has which tools
Inside one small business this gap is easy to miss, because nobody decides it on purpose. Five questions make it visible.
- List every AI tool in use. Free and paid, including the AI features built into email, office and accounting software.
- Mark who pays for each one. The business, one person out of pocket, or nobody.
- Write down who holds each paid seat. Then ask whether that was a decision or something that just happened.
- Note what the free tier cannot do. Longer files, uploads, connections to email or calendar, agent features. That list is where the gap now sits.
- Choose one person and one task for a paid trial. The step below shows how to judge it within a month.
The one-month test, in one line
Hours the task took last month, minus hours it took with the paid tool, times what an hour of that person’s time is worth. If the answer is more than the subscription, keep it. If not, cancel before the second month.
One small thing, and it is not a course
Take the task that ate the most hours last month. Price one month of the paid tier of one tool, run that task through it, and answer one question honestly: did it earn the subscription back. One task, one month, one number from this business rather than an average from somebody else’s. Much smaller than a training program, and it needs less space to learn than most owners assume they need.
One more worth writing down, if anyone else works in the business: who holds the paid seat. That list is usually short, and almost never deliberate.
The adoption charts will keep converging, and keep being quoted as progress. The tool-by-tool numbers suggest the gap relocated to where the real capability lives, and capability now arrives with a price and a license attached. Which leaves a question no usage chart answers: when the tools that actually do the work have to be bought and granted, who decides which women get them?
More research on this subject is collected on the Women in AI topic page.
Related guide: Free AI Courses for Women, Checked November 2026, twelve programs confirmed on their official pages.
Sources
The facts in this article were checked against these pages on November 12, 2026.
In this article

Your AI Skills Live in One Head. Here Is the Fix
Adoption hit 86% but just 15% fully integrated AI. What the small business AI skills gap really is, and one step to start closing it this week.
Looking for something else?