Houston, we have an AI problem: the Gender AI Gap

Previous waves of automation primarily affected factories. The current AI revolution is transforming desk work – and women are being disproportionately disadvantaged in the process. What lies behind the Gender AI Gap, which three findings are particularly surprising, and what can companies do about it?

AI for everyone – except women?

Whether at work or in private life, more and more people are using artificial intelligence, arguably the most important technological innovation of the decade. The problem is that men are using it more. Women use AI less frequently and less intensively in their professional lives – a phenomenon that is already well documented and has become known as the “Gender AI Gap”.

This is no longer a niche issue. It is a tangible productivity and equality problem that grows month by month – in skills and career opportunities – much as we once observed with the Gender Pay Gap.

Earlier waves of automation primarily affected physical work in male-dominated factories. Generative AI turns that pattern on its head: it changes cognitive desk work – analysis, coding, emails, concepts and much more. And it is precisely here that a new, empirically documented gap between the sexes is emerging: a gap in access, usage, skills and attitudes – and, particularly concerning, a promotion gap if women take on strategic AI roles less frequently.

Importantly, this gap is not an expression of lower AI capability among women. It is the result of structural conditions – employer support, perceptions of risk, unequal networks and different time budgets. That also contains the good news: what has emerged structurally can also be closed structurally.

Sixteen percentage points that matter

Current figures illustrate the scale of the gap. AI has long since entered everyday working life. According to Statista, 77 per cent of employees worldwide used AI at work in 2025, with adoption even higher among younger people: 84 per cent of 18- to 34-year-olds compared with 69 per cent of 35- to 54-year-olds. The gender perspective is particularly revealing: worldwide, women are more strongly affected than men by the automation effects of generative AI, especially in highly digitalised regions.

The gap in AI use tends to be particularly pronounced in technology-intensive sectors such as IT, engineering and financial services, while it is smaller in social and creative professions.

For Germany, the IAB study “Digital Gender Gap – Focus 2026: Artificial Intelligence” provides the most detailed picture to date. It identifies an AI usage gap of around 16 per cent between women and men of working age. The study's authors argue that the gap is less an expression of individual reluctance than the result of social expectations, organisational structures and unequal opportunities to use the technology.

Three surprising findings

The Gender AI Gap appears – to varying degrees – almost everywhere it has been measured so far. Anyone assuming that the problem will simply disappear with the next generation may be surprised by the current evidence.

Surprise 1 – younger people are not doing better. According to the study mentioned above, in “Generation Z+” – roughly those born between 1996 and 2010 – around one in two young men already uses AI intensively, compared with fewer than one in three young women. The gap is therefore larger here than among older employees.

Surprise 2 – education and income alone do not solve the problem. AI use does increase with higher educational attainment. However, men benefit from this more strongly than women. Informal networks often reproduce existing inequalities rather than reducing them.

Surprise 3 – the resilience paradox. Among men, a relaxed attitude towards technological change has no measurable effect on AI use. Among women, the picture is reversed: particularly resilient women actually use AI less frequently than less resilient women.

A brief look at the causes

Where does this gap in AI use come from? The causes are complex and cumulative:

Confidence gap: Men systematically rate their AI knowledge more highly than women do – regardless of their actual level of competence. Women more frequently express concerns about misinformation, data privacy and dependency.

Risk perception with a rational basis: On average, women display greater risk aversion, including towards economic disruption – and they are in fact more frequently employed in occupations that are vulnerable to automation.

Role expectations and missing role models: Stereotypes about lower affinity for technology weaken confidence and willingness to use new tools. This is reinforced by a lack of visible female role models and the underrepresentation of women in STEM subjects and AI training.

Imbalance at executive level: Decisions about AI budgets and software procurement are generally made where women – as in traditional IT development – remain underrepresented. As a result, female perspectives are less likely to influence how technology is designed and introduced.

Why there is no time to lose

These causes have tangible consequences. Where women remain excluded from AI applications, companies incur real efficiency losses and innovation gaps. At an individual level, lower AI use can mean fewer opportunities for promotion and salary growth. The urgency comes from the speed of AI-driven social change: in just four years, the technology has moved from the margins to become a significant productivity factor. Those who start using AI early and regularly accumulate experience faster. Those who enter later first have to overcome a structural head start that others have already built.

What companies can do now

The good news is that the Gender AI Gap can be changed. The most effective lever lies within the company itself – above all through employer-funded training. Concrete measures include:

Protect roles with high exposure: Where women work in areas particularly affected by automation, prioritise upskilling, process redesign and employee participation.

Create equal access: Introduce mandatory, role-specific training paths.

Actively manage AI adoption: AI should be introduced systematically for everyone rather than leaving adoption to chance.

Talk about benefits rather than fear: Internal communication should emphasise the value AI can create in everyday work rather than relying on fear of job losses.

Build trust and guardrails: Establish clear policies for use cases, data privacy and responsibilities, complemented by internal audits and easily accessible support.

Design training with gender differences in mind: Tailor learning opportunities to different starting points rather than relying on one-size-fits-all solutions.

Over the longer term, another important lever is early STEM support for girls and young women – from school through to continuing professional development.

The Gender AI Gap is not only a warning signal; it is also an invitation to rethink training, leadership culture and tool design more inclusively. If roughly half of the workforce systematically uses a key technology less frequently, that is simply untapped productivity potential – your productivity potential. And what does the situation look like in your company?