The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Women are entering artificial intelligence, but they remain underrepresented in its workforce and research roles, and gaps persist in leadership. That matters because decisions about AI’s design, deployment and rules are made by people with different levels of access and influence. The evidence shows both gradual change and substantial barriers; it does not show that women share one perspective or that representation alone makes AI fair.
How underrepresented are women in AI?
UNESCO’s leadership article, published on 11 December 2024 and updated on 17 April 2026, estimates that women make up 30% of AI professionals. That is a useful signal of imbalance, not a universal census: the article does not provide one standardized definition of who counts as an “AI professional.”
UNESCO’s 2024 Women for Ethical AI Outlook Study compares several workforce and leadership indicators. These figures describe gaps from parity, not women’s share of a workforce. They also come from different roles and contexts, so they should not be treated as a single dataset or as steps in a measured promotion funnel.
| Indicator | Reported gender gap | How to read it |
|---|---|---|
| Science research and development positions | 21% | Gap from parity, according to UNESCO’s 2024 Outlook Study. |
| AI research positions | 38% | Gap from parity; this is a research-role indicator, not the share of all AI workers who are women. |
| ICT professionals | 15% | Gap from parity; definitions and source populations vary across the study’s indicators. |
| Software development professionals | 44% | Gap from parity; not directly interchangeable with the ICT figure. |
| Director positions at STEM workplaces | 23% | Leadership indicator from STEM workplaces. |
| C-suite positions at AI startups | 32% | Leadership indicator from AI startups; not a like-for-like comparison with STEM directors. |
The pattern is the important point: UNESCO’s selected indicators show wider gaps in some AI-specific roles than in broader comparison categories. But because the measures differ in role, population and definition, they cannot establish exactly where or why women leave the path to AI leadership.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
Is the picture changing?
Yes, although growth is not the same as parity. The World Economic Forum’s 2024 report, using LinkedIn member data from 166 economies, says women’s share among AI talent increased over the previous four years while men remained substantially more represented. LinkedIn profiles capture only part of the labor market, so the trend should be understood as a change in the platform’s data rather than a census of every AI worker.
UNESCO’s 2024 leadership article also reports that about 37% of AI inventors named on patents filed in 2022–23 were women. The wording matters: this is a figure about inventors named on patents during that period, not a claim that women owned 37% of AI patents.
Rank #2
Where does influence over AI get decided?
Influence is not limited to executive titles. It can arise at several points: who gets the education and technical opportunities to enter the field; who is hired, supported and promoted; and who has a say in the design, deployment and governance of AI systems. UNESCO’s indicators speak most directly to workforce representation and selected leadership roles. They do not, by themselves, measure who has influence over each product or policy decision.
That distinction helps explain why a headcount alone is an incomplete measure. More women in technical jobs or leadership may broaden whose questions and concerns are present, but women are not a uniform group, and representation by itself does not guarantee a fair system. Claims about a particular system or team need evidence about that system and its development, not an assumption about the gender of its creators.
What does representation have to do with AI outputs?
Representation is one reason to scrutinize how systems are designed and evaluated, but the connection should not be reduced to a claim that a team’s gender composition caused a model’s output. UNESCO’s 2024 summary of its study reports that, in stories generated by Llama 2 under the study’s prompts, women were described in domestic roles four times more often than men. That finding is specific to the tested model and prompts; it is not a measurement of every model or all generated text.
The result makes evaluation concrete: researchers and organizations can examine what a model produces under specified tests and monitor for recurring stereotypes. It does not establish that any single hiring change would prevent those outputs. UNESCO’s Director-General communication calls for governments to establish clear regulatory frameworks and private companies to carry out continuous monitoring and evaluation for systemic bias, referring to the UNESCO Recommendation on the Ethics of Artificial Intelligence adopted by member states in November 2021.
What responses does UNESCO recommend?
The 2024 Women for Ethical AI Outlook Study calls for better evidence and action across AI’s design, use and governance. Its recommendation is: “It emphasizes the importance of collecting more comprehensive and disaggregated data, implementing targeted interventions, and developing inclusive policies that promote equitable participation in the design, use, and governance of AI.”
- Improve measurement: Collect more complete, gender-disaggregated data so gaps can be tracked across roles and stages rather than inferred from mismatched indicators.
- Use targeted interventions: Address identified barriers with measures suited to the populations and settings involved, rather than assuming one program will fit everyone.
- Make policy inclusive: Include equitable participation in decisions about AI’s design, use and governance.
- Support career development: UNESCO describes the Organization for Women in Science in the Developing World as offering research training, career development and networking opportunities to women scientists at different career stages. It is an adjacent science-support resource, not evidence of an AI-specific service.
What the available figures can—and cannot—tell us
The institutional evidence establishes continuing underrepresentation and points to gaps in research, software development and selected leadership categories. It also records signs of change in LinkedIn’s AI-talent data and a meaningful share of women among named inventors on AI patents filed in 2022–23. The figures do not reveal every worker’s experience, explain each career decision, or establish a universal causal link between workforce composition and AI outcomes.
Best Value
Understanding who gets to leave a mark on AI therefore requires more than one percentage. It requires comparable data on entry, retention and leadership, plus careful scrutiny of the systems and rules being built. The evidence supports treating equitable participation as a serious workforce and governance issue—not as a shortcut for predicting what any individual woman will think or what any AI system will do.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
