On Monday 12 October, techUK hosted a Women in Tech Group breakfast event, Closing the gap: Women, work & AI adoption, to open Women in Tech Week and mark the day before Ada Lovelace Day. The event brought together perspectives from research, industry, and policy to explore how women are being affected by the development and use of AI at work, and what employers and government can do to make sure the benefits reach the whole workforce. 

Opening the event, Sheila Flavell CBE noted that women make up around 22% of the UK’s AI workforce and adopt AI tools at a rate roughly 25% lower than men, risking a vicious circle that widens existing pay and progression gaps. She argued that progress should be measured by how many people benefit from AI, not just how many use it. We were pleased to be joined by: 

• Sheila Flavell CBE, President, techUK and Chief Operating Officer, FDM Group 

• Rachel Seftel, Head of Research and Impact, Cherie Blair Foundation for Women 

• Sue Daley OBE, Director of Technology and Innovation, techUK 

• Maryam Yaqub, AI Lead, Fawcett Society 

• Riham Satti, CEO and Co-founder, MeVitae 

• Cat Pinchard, Head of Future of Work and AI Enablement, Lloyds Banking Group 

Event overview

In-conversation: Adopted, not embedded

Rachel Seftel presented the Cherie Blair Foundation for Women’s report, Adopted Not Embedded, a survey of over 3,000 women entrepreneurs in low and middle income countries. AI use among respondents more than doubled in a year, from 38% to 82%. As Rachel put it, the adoption train “left the station an hour ago” as the question now is how deeply women embed AI, and who captures the value. 

Women using AI in back-end functions such as finance and operations were 1.8 times more likely to report increased revenue than those using it only in customer-facing tasks. Time savings were the most cited benefit, but Rachel cautioned that, given women’s disproportionate caring responsibilities, time saved is not always time reinvested in the business, it may go to school pickup or sleep instead. 

Only around a third of women surveyed fully trust AI, yet most use it anyway, having judged the time savings worth the risk. Rachel argued that this caution is rational and grounded in real risks, not a lack of confidence. The answer, she argued, is to tackle the causes of that caution rather than ask women to set it aside: “We shouldn’t be telling women that they need to worry less. We should be giving them less to worry about.” That puts the onus on those building and deploying AI – as she noted, if customers aren’t buying what you’re selling, the problem is the product, not the customer.

Barriers to gaining value from AI

The Fawcett Society’s research with the Open University shows women-dominated roles face three times the AI exposure of male-dominated ones, across both low-paid and high-paid work. Maryam Yaqub pointed to structural barriers rather than a confidence gap. Women are less encouraged by managers to use AI, less recognised when they do, and more likely to be seen as cutting corners; in her words, women face “real costs that men don’t absorb”. 

Those costs are compounded by a lack of organisational guidance. Only around 17% of UK businesses have an AI policy, guidance or strategy in place, which leaves many women weighing legitimate safety and ethical concerns, and the reputational risk of being seen to use AI, without clear rules or backing from their employer. Maryam argued that clear policies, combined with training tailored to specific roles rather than generic prompt engineering, help women see where AI reduces their workload instead of adding to it. She also urged employers to be honest upfront about job displacement and to create pathways for women to retrain. 

Riham Satti framed the barrier as one of trust, noting that with AI “we get an output, we don’t always get the working out”. Cat Pinchard shared that 82% of Lloyds Banking Group colleagues were positive about AI use in its latest survey, with less confident staff asking for role-relevant use cases and clear guidance on the do’s and don’ts.

Bringing the whole workforce along 

Panellists shared approaches including pairing early adopters with more cautious colleagues and starting from the problem rather than the technology (MeVitae), training leaders first and then offering persona-based learning through an AI Academy (Lloyds) and starting with data quality and a network of AI champions rather than pilots (Basildon Borough Council). 

Human oversight and bias 

Riham warned that “human in the loop doesn’t mean that a human needs to intervene”, arguing that the term should give way to ‘human oversight’, with a named person accountable for automated decisions. Maryam noted that equality action plans are expected to become compulsory from spring 2027 and should cover AI across the whole employment life cycle, adding that women cannot challenge a decision if they are not aware AI was used to make it. She cited an LSE experiment in which identical social care notes produced different summaries when only the gender was changed.

Key takeaways 

• Measure benefit, not usage: Data should be disaggregated by gender so that disparities hidden in headline adoption figures can be addressed. 

• Depth drives value: The biggest gains come from embedding AI in core business functions, supported by problem-led, role-specific training. 

• Women’s caution is rational: Safer products, clear AI policies and meaningful human oversight will do more than asking women to trust AI more. 

• Representation matters: With women making up around 22% of the AI workforce, gaps in who builds AI show up in biased outputs. 

Watch the full event recording 

If you missed the event, you can watch the recording below: 

Get involved

If you're a techUK member interested in getting involved in our women in tech workstreams, join our Women in Tech Group, a network dedicated to supporting women at every stage of their tech careers. For more information, please contact Tess (Tess.newton@techUK.org).