From leaky pipelines to launchpads: why diversity is the missing link in tech’s AI revolution

To name just a few from the sector I work in, this could include climate change, security across the food-energy-water nexus and preservation of biodiversity. The eruption of AI provides a pivotal moment to reset, focus on the big challenges and accelerate to a brighter future.  

To make the most of this ‘moment to reset’ we need to get our house in order. Whilst the technology sector has long been synonymous with innovation, it lags behind when it comes to diversity. But why does that matter? 

Diversity is a catalyst for innovation: ‘In the knowledge economy, success depends on bringing in diverse perspectives and making use of them’ⁱ. It’s important to say early on that diversity is a broad church, including gender, race, sex, age, religion, ethnicity, disability, socio-economic and other dimensions. Gender is often considered a tractable dimension of diversity, so is sometimes used as a proxy for other aspects. I like to think (hope) that steps to address gender diversity will deliver benefits that ripple across other dimensions: "A rising tide lifts all boats."  

So, I was keen to read the Lovelace 2025 reportⁱⁱ published in July 2025, and was disappointed to read that just within the UK just 20% of the tech workforce are women. This figure poignantly echoes the 2021 ‘Where are the Women?’ⁱⁱⁱ policy briefing that also reported that 20% of the professionals working in AI and Data Science in the UK were women. Given the excitement around emerging tech, such as AI, has the needle really not shifted in the past 4 years? 

The tech workforce lacks the rich diversity needed to supercharge innovation. This is a problem. As Professor Dame Wendy Hall, UK AI Council Skills Champion and author of the 2017 review ‘Growing the Artificial Intelligence in the UK’^iv, said ‘If it isn’t diverse, it isn’t ethical’. That should be enough to trigger action. But if it’s not, then there is an abundance of evidence showing that diverse organisations are more successful when recruiting and retaining talent^v, have better financial performance^vi, higher performing teams^vii and a broad range of other benefits^viii. Diversity catalyses innovation^ix, at a time when we need it most.  

So what can be done?  It’s hard to know where to start, but start we must. The Lovelace report estimates that the loss of women from the tech sector costs the UK between £2 billion and £3.5 billion every year. This would be concerning at any stage, but we have ambitious plans for kickstarting economic growth in the UK^x and need all hands on deck.  

As the AI Opportunities Plan^xi highlights, achieving gender parity ‘would mean thousands of additional workers’ if women could be encouraged to enter, or remain in, the sector. And there’s more, AI reflects its creators and source (training) data. A lack of diversity in those collecting the data, or creating AI tools, risks a future where AI tools do not represent, or serve, the needs of the full breadth of society. The repercussions of which are well described by Caroline Criado Perez^xii.  

Systemic change is needed to stem the ‘leaky pipeline’ as young girls leave STEM subjects, to ‘dismantle the barriers’ facing women in research^xiii and, at an organisational level, to embed thinking about how careers are ‘truly built, supported, and sustained’^xiv.   I’ve seen this fail, but I’ve also seen it work.  

At the Met Office ‘We believe great minds don’t think alike, and diversity is a strength. Differences in background, lived experience, culture and perspective can enrich our working environment and help us deliver world-changing work'^xv.  

I see this in action, and that’s what’s needed: action. At the Met Office we’ve taken steps to strengthen networks, especially those for minority or marginalised groups. These are safe spaces where employees with a shared identity or life experience can find a sense of community and allyship. We also mark the year with events to celebrate diversity. For example, we hold an annual event to celebrate Ada Lovelace Day (a cause close to my own heart). We’re not there yet, but our journey is underway and we’re taking significant (and measurable) steps towards achieving an inclusive workforce, where employees feel valued, respected and fairly treated.   

And what for us as individuals? If you’re a woman working in tech, I urge you to consider whether you’re visible. It might feel uncomfortable, but it is important. ‘You have to see it to be it’ and you never know who may be looking for a sign that a future career in tech is possible.  

We should all encourage any young person with an interest in a STEM^xvi subject to stick with it. There is an alarming and persistent tendency to stereotype subjects as being more suited to one gender than another. This is wrong. It does a disservice to the next generation, who should be nurtured and encouraged to be the best they can, in whatever subject they choose. Now is the time to encourage young people to study STEM subjects- not leave! As before, it’s ‘all hands on deck’.  

Finally, some research cites a concern that ‘Data science can feel abstract and without purpose’^xvii. This one cuts to the quick. We must address data science’s image problem. Far from ‘without purpose’, data science, and other technologies, may hold the solution to some of the gravest challenges facing society, such as climate change and food security.   

Creating a diverse technology workforce can’t be a one-off initiative - it also can’t be left to governments, organisations or (other) individuals. We will all need to play our part. It’s a sustained, systemic effort that requires focus, drive, data-driven monitoring, inclusive policies, cultures that celebrates difference as a strength, and individuals who are willing to live by the values of inclusivity and diversity. Embedding and celebrating diversity in the fabric of our workforce and society will be essential to shaping a more equitable, effective and more positive future. 

I write this hoping that we’re building change that will ripple out; let’s embed diversity in the fabric of our tech workforce. I mean diversity in the broadest sense: including, but not limited to the usual metrics such as gender, race, age and religion.  

We could now strive for inclusion ‘beyond the metrics’ and take steps to also include a range of skills, backgrounds, experience, socio-economic background, education and disciplines. It’s in this diversity that the magic really lies. Let’s act now to create the rich, fruitful, challenging, innovative, productive – and ultimately rewarding workforce that we need for today and generations to come.

References

_{i Solheim, M.C., 2022. Diversity and inclusion is a must to make innovation work for all. Nature, 612(7939), pp.11–11.} 

_{ii https://wearetechwomen.com/the-lovelace-report-wearetechwomen-oliver-wyman-reveal-barriers-facing-women-in-tech/} 

_{iii Young, E., Wajcman, J. and Sprejer, L. (2021). Where are the Women? Mapping the Gender Job Gap in AI. Policy Briefing: Summary. The Alan Turing Institute.} 

_{iv https://www.gov.uk/government/publications/growing-the-artificial-intelligence-industry-in-the-uk} 

_{v Madera, J.M., Ng, L., Sundermann, J.M. and Hebl, M., 2019. Top management gender diversity and organizational attraction: When and why it matters. Archives of Scientific Psychology, 7(1), p.90.} 

_{vi Carter, N.M. and Wagner, H.M., 2011. The bottom line: Corporate performance and women's representation on boards (2004–2008). Catalyst, 1, pp.1–3.} 

_{vii Shoreibah, R.A., Marshall, G.W. and Gassenheimer, J.B., 2019. Toward a framework for mixed-gender selling teams and the impact of increased female presence on team performance: Thought development and propositions. Industrial Marketing Management, 77, pp.4–12.} 

_{viii Catalyst, Quick take: Why diversity and inclusion matter (June 24, 2020).} 

_{ix https://www.bcg.com/publications/2017/people-organization-leadership-talent-innovation-through-diversity-mix-that-matters} 

_{x https://www.gov.uk/missions} 

_{xi https://www.gov.uk/government/publications/ai-opportunities-action-plan/ai-opportunities-action-plan} 

_{xii Perez, C.C., 2019. Invisible women: the Sunday Times number one bestseller exposing the gender bias women face every day. Random House.} 

_{xiii Jebsen, J.M., Nicoll Baines, K., Oliver, R.A. and Jayasinghe, I., 2022. Dismantling barriers faced by women in STEM. Nature Chemistry, 14(11), pp.1203–1206.} 

_{xiv https://wearetechwomen.com/the-lovelace-report-wearetechwomen-oliver-wyman-reveal-barriers-facing-women-in-tech/} 

_{xv https://www.metoffice.gov.uk/about-us/who-we-are/our-values/equality-diversity-and-inclusion#:~:text=Pay%20Gap%20Report-,Our%20approach,us%20deliver%20world%2Dchanging%20work.} 

_{xvi Science, Technology, Engineering, and Mathematics}

_{xvii Duranton, S., Erlebach, J., Brégé, C., Danziger, J., Gallego, A. and Pauly, M., 2020. What’s Keeping Women out of Data Science?. BCG [online].}