Tackling the gender skills gap in AI not only addresses the moral imperative of inclusivity, it ensures stereotypes and outdated thinking are eliminated at source, says Petra Jenner of Splunk
The European Commission’s latest ‘State of the Digital Decade’ report revealed that most member states are way off target to meet their 2030 digital transformation goals.
The bloc’s Digital Decade programme sets key targets in areas such as digital skills, infrastructure, and AI adoption to boost the region’s global competitiveness. However, with only 55.6% of the EU population possessing basic digital skills and a projected AI adoption rate from businesses at 17% by 2030, well below its 75% target, Europe risks falling short. To achieve its ambitions for prosperity and growth, The EU needs bold, inclusive transformation.
As the stakes mount, generative AI (Gen AI) and more “traditional” applications of machine learning (ML) technology offer enormous potential to boost productivity. McKinsey research suggests that, as the use of Gen AI in the workplace grows, it could unlock up to €4tn for the global economy each year. Europe’s success in capturing its share will hinge on its ability to develop responsible AI that enhances human potential and reflects societal needs.
Building an inclusive foundation
Placing people at the centre of societies’ digital transformation is the core thinking behind the Digital Decade policy programme. This means ensuring that a pivotal technology like AI represents everyone.
However, the underrepresentation of women in the development of AI systems is a glaring issue. With women making up just 22% of AI professionals globally, as reported by the World Economic Forum (WEF), there is a real risk that AI development will perpetuate existing biases, creating untrustworthy systems that fail to serve society equitably.
To address this, it is crucial to focus on enabling women to better utilise AI in the workplace. Despite AI’s potential to help bridge the gender gap in technology and beyond, many women are not receiving the necessary training or support to take full advantage of these tools. This gap is especially concerning given that women are overrepresented in roles most likely to be impacted by automation, with 80% of such jobs at risk, according to US data from the Kenan Institute.
By providing targeted training and resources, we can ensure that women are not left behind in this digital transformation, but instead, are empowered to thrive in an AI-driven future.
Turning lip service into action
The first critical layer of bias begins with the people at the table when AI systems are being developed. If the teams behind these technologies are not diverse, the AI systems they build will likely mirror their creators’ biased perspectives.
As an industry, we must stop paying lip service to gender equality, and instead fundamentally rethink how to encourage greater participation. Crucially, this effort should begin before people enter the world of work, with meaningful school outreach to encourage more girls and young women into STEM careers.
In the workplace more generally, meanwhile, we must ensure that job opportunities and advancements are accessible to all. There is a leaky pipeline when it comes to working in tech. According to Accenture, whereas over half of male computer science graduates stay in the field, only 38% of female graduates do. And half of all women in tech roles leave before they are 35 years old.
Job adverts, for example, should be written and promoted with a focus on inclusivity and the wider skills needed to drive successful AI adoption. Visible female role models will also be critical to more young women seeing AI (and other tech routes) as a viable career path.
Ultimately, diverse representation is not just a matter of ensuring fairness in the workplace (though that, of course, is critical too). It’s also about ensuring the quality and reliability of AI – one of our most critical tools today. Otherwise, AI is at risk of reinforcing existing stereotypes and inequities, undermining its potential to benefit society.
Proactively addressing bias
AI systems do not, after all, create information, but rather reflect human thoughts and data, and the biases of the internet. As AI adoption expands into sectors like banking, insurance, healthcare, and HR, these biases could threaten people’s finances, jobs, and lifestyles, entrenching discrimination and eroding trust.
To avoid perpetuating outdated thinking, decisive action is needed. A recent study by the Berkeley Haas Center for Equity, Gender and Leadership found that almost half (44%) of the AI systems they investigated showed gender bias. Another UNESCO study highlighted how Large Language Models (LLMs) tend to reproduce gender biases via language scraped from the internet, often associating women in domestic roles with words like “home”, “family” and “children” while linking men to terms like “business”, “executive”, “salary”, and “career”.
Eliminating these biases is no small task, but as the EU takes the lead on AI regulation with the ambitious EU AI Act, it’s vital that we get it right. Yes, as provided by the AI Act (Article 10 on Data Governance), we must ensure AI training datasets are diverse and representative. And yes, systems and algorithms should be regularly audited to address discriminatory outcomes. However, we also need to think bigger. Dedicated initiatives are urgently needed, especially among companies developing LLMs. We need committees focused on identifying and combating gender bias, potentially even using AI to aid in detection. A gender bias index could help companies choose higher-performing AI toolsets, while independent analysts could provide gender rankings of LLMs.
Bridging the skills gap
As we work to eliminate biases in AI, we must also address the broader challenge of equipping the workforce for the AI-driven future.
Studies show that female colleagues are more likely to feel under equipped to flourish in the AI age. The Female Quotient’s ‘State of AI + Women’ report released at Davos 2024, for example, shows that women are 20% more likely than men to feel behind the curve.
Beyond being the right thing to do, building female employees’ AI skills makes clear business sense. A diverse and well-trained workforce brings different perspectives, which enhances creativity, problem-solving, and the ability to identify and mitigate risks – leading to more robust, competitive, and resilient organisations.
Trust in the future
The future of AI is incredibly bright, but its potential could be undermined if we fail to build trust in its outcomes. That’s why we must ensure that AI reflects all aspects of the human experience. Those at the forefront of the AI revolution bear a significant responsibility: to create a technology landscape where everyone can contribute to shaping the future. Inclusivity isn’t just a moral imperative – it’s essential for calming AI fears and ensuring that technology complements, rather than replaces, human decision-making.
Ultimately, bringing in employees from a wider, more representative spectrum will ensure that AI systems are developed and driven in a way that benefits organisations, their people, and society as a whole – in Europe and beyond.
About the author
Petra Jenner is Senior Vice President, and General Manager, EMEA at Splunk.