Artificial intelligence is already shaping decisions about pay, health, public services and online safety. Without effective measures in place, these systems risk embedding existing gender inequalities into automated systems operating at scale.

Our white paper examines how AI may affect women’s economic equality and the wider systems that shape women’s lives, including education, healthcare, policing and welfare. It sets out the steps policymakers, regulators and organisations must take to ensure innovation supports equality rather than reinforcing existing disparities.

⬇ Download White Paper

⬇ Download One-Page Summary

---

Quick Stats

• Over 95% of deepfake content online is pornographic, overwhelmingly targeting women.
• Automation and AI could displace up to 40% of women in certain industries by 2030.
• 50% of online men vs 33% of online women have tried generative AI.
• Women make up only around 22% of the UK AI workforce.
• AI salary tools have been shown to recommend lower pay to women than to identically qualified men.
• Over 80% of documented AI failures were driven by privacy violations or algorithmic bias.
• 72% of the UK public say laws and regulation would increase their comfort with AI technologies.

Where Harm Is Emerging

Employment AI tools increasingly shape recruitment, performance monitoring and workforce planning. Systems trained on historically unequal labour-market data risk replicating gender pay disparities and occupational segregation.

Healthcare Clinical AI systems are often trained on male-dominant datasets, which can contribute to underdiagnosis and misclassification of women’s health conditions.

Public services Automated risk scoring systems influence decisions about welfare eligibility, policing and social services, potentially reinforcing existing patterns of disadvantage.

Digital spaces Generative AI has accelerated the creation of deepfake imagery and other forms of image-based abuse that disproportionately target women.

Why AI Bias Is Different

Artificial intelligence has the potential to embed existing inequalities into infrastructure.

Scale
Automated systems can influence thousands of decisions simultaneously.

Speed
Algorithmic outputs are produced instantly and continuously.

Systematisation
Bias becomes embedded within institutional systems and workflows, often appearing neutral or objective.

These characteristics mean that small distortions in data or system design can produce large-scale inequality.

 A Framework for Bias-Resistant AI

The research proposes six principles for designing AI systems that advance equality rather than entrench inequality:

• Equality by design
• Representative data
• Diverse teams
• Transparency
• Accountability
• Empowerment

Together, these principles treat equality as a system requirement across the entire AI lifecycle, from commissioning and procurement to deployment and oversight.

 Key Policy Recommendations

The report calls for stronger safeguards to ensure AI systems are developed and deployed responsibly.

Key recommendations include:

• Introduce cross-sector AI legislation establishing statutory duties for high-risk systems
• Require equality impact assessments before high-risk AI systems are deployed
• Establish a national register of high-risk AI systems
• Strengthen transparency and contestability where AI affects employment, pay or services
• Introduce safeguards against AI-enabled sexual image abuse

⬇ Download the full report 

⬇ Download the one page summary 

 Explore key Insights

We have developed a series of explainers and visual summaries highlighting the main findings from the research. 

Follow us on Instagram to see more

About the Research

This white paper forms part of the Fawcett Society’s work examining the relationship between artificial intelligence and gender equality.

Further analysis is currently being undertaken through a collaboration with Responsible AI UK, which is exploring how AI adoption may affect the gender pay gap and wider labour-market outcomes.

AI is already shaping economic opportunity. Without deliberate and thought-through safeguards, it risks embedding gender inequality into the systems that govern employment and public services. The detail is technical, the consequences are not - we cannot allow women's financial equality be harmed by technological advancement. AI if implemented carefully at employer level could root out inequalities, but only with safeguards, knowledge and regulation. At Fawcett, we exist to root out bias and sexism. Whether this comes from people, systems or technology. We will not be quiet as women's opportunities and access is impacted by AI being implemented without due care and attention to the consequences

Penny East, CEO

AI does not create inequality from nothing. It absorbs the biases already present in our institutions and reproduces them at scale. If we fail to intervene, systems that already disadvantage women will simply be automated

Maryam Yaqub, Author

Research collaboration: AI, skills and the gender pay gap

Alongside this white paper, the Fawcett Society participated in a research project funded by Responsible AI UK, led by the Centre for Protecting Women Online at the Knowledge Media Institute, The Open University, and delivered in partnership with Politecnico di Torino (Italy) and Nokia Bell Labs.

The project examined how AI innovation interacts with occupational tasks, skills and pay structures, with a particular focus on whether emerging technologies may amplify or reduce gender inequalities in the labour market. By linking data on AI innovation with occupational and income data, the research aimed to build stronger evidence on how AI adoption may influence pay outcomes, job design and skills demand across different sectors.

A key strand of the research explored AI adoption and skills gaps, including whether unequal access to AI-related skills and training may create new productivity and progression gaps between men and women.

The findings from this research were published in a standalone report and informed the analysis presented in this white paper, strengthening the evidence base on the relationship between AI adoption, skills and gender pay inequality.

⬇ Download the Responsible AI UK research report

AI and Gender Equality Roundtable: Bringing the conversation to Parliament 

In April 2026, the Fawcett Society convened an invite-only policy roundtable at the House of Lords, in collaboration with the Misogyny and AI Network. The event brought together Members of Parliament and peers from the House of Lords, alongside barristers, trade unions, researchers, civil servants, regulators and leaders from across the women’s sector.

• Find out more here

Share This Research

Help spread the conversation on AI and gender equality.

Stay Involved

Join the Fawcett Society supporter mailing list to stay informed about research, campaigns and opportunities to take action on gender equality.

• Join the Supporter Mailing List