The International Labour Organization (ILO) reports that globally, one in four workers are in occupations with some GenAI exposure, with the transformation of jobs the most likely impact. Yet research shows that women have approximately 22% lower odds of using GenAI tools than men. Compounding this, Deloitte’s researchers found that only 61% of women GenAI users feel their company actively encourages use of the technology at work, compared with 83% of men, while only 49% of women report that their company invests in GenAI training, compared with 79% of men.

The implications are significant. Businesses risk forfeiting productivity gains, embedding gender bias into AI systems, and – for designated employers – undermining the transformation objectives they are legally required to pursue. Understanding these risks is the first step toward addressing them.

The commercial concern

For businesses, the most immediate concern is lost productivity. Properly deployed, GenAI tools can deliver significant efficiency gains – one study reports GenAI users saving an average of 2.2 hours per week over a 40-hour workweek. If women are not adopting these tools at comparable rates, their disengagement costs businesses time and money.

Lower adoption also leads to inefficiencies. If women remain underrepresented as GenAI users, the AI systems themselves will become progressively less effective for tasks women perform and less reflective of women’s needs and preferences. Where a predominantly male user base produces training data or provides feedback on outputs, there is an inadequate sample of women’s perspectives when it comes to training AI and critiquing its outputs. Research shows that as many as 44% of AI systems across industries reflect gender bias, which can reduce the utility of GenAI for women, creating a feedback loop that disincentivises engagement.

Bias perpetuated through GenAI can also lead to unfair discrimination against women. Consider the instances of women being discriminated against during hiring processes or facing reduced access to financial services. For South African businesses deploying GenAI in recruitment, performance management, customer engagement, or lending, these biases create legal, operational and reputational risks.

Some studies suggest that women are less likely than men to use GenAI tools due to heightened risk awareness such as concerns about an absence of clear oversight, data privacy and a lack of accountability mechanisms. While this research is not conclusive, a reduction in advocates for a methodical approach to AI adoption can weaken businesses’ strategies around technological deployment, leading to implementation without adequate recognition of risk.

The exposure paradox

The ILO suggests that female-dominated occupations – such as business administration, clerical support and secretarial roles – are almost twice as likely to be exposed to GenAI as male-dominated occupations, and that people in these roles face much higher automation risk. Paradoxically, this may mean that women are both most exposed to GenAI disruption and least likely to be adopting it.

Lower and slower adoption of GenAI by women risks widening the gender pay gap and reducing promotion opportunities, as productivity gains accrue disproportionately to men.

This is not a matter of capability and is not limited to female-dominated occupations. In 2025, a study of academic researchers found that, following the emergence of ChatGPT, the increase in male researchers’ productivity was 6.4% higher than that of female researchers, and the productivity gap between male and female researchers widened by 57.1%. Applied across industries, such abrupt shifts can materially influence career trajectories, especially when compounded over time.

The compliance imperative

The EEA requires designated employers, those with 50 or more employees, to conduct a detailed analysis to identify employment barriers that adversely affect people from designated groups, which include Black people, women and people with disabilities.

An employment barrier is any policy, procedure, practice or condition that limits or prevents equitable access to employment opportunities for designated groups. This analysis forms the foundation of the employment equity plan (EE plan) that designated employers must prepare and implement.

Given the research outlined above, unequal access to GenAI tools and training, where women are less likely to receive encouragement or investment in GenAI skills, may constitute precisely the kind of employment barrier that designated employers are obliged to identify. If women are systematically excluded from productivity enhancing technologies, or if workplace cultures discourage their adoption by women, this can limit equitable access to career advancement opportunities.

The compliance imperative is therefore clear: designated employers conducting their barrier analysis should consider whether disparities in GenAI adoption and training constitute an employment barrier affecting women and, if so, include measures to address this in their EE plans.

The strategic response

Designated employers are already required to prepare and implement EE plans that address identified employment barriers. They can adapt this existing framework to tackle the GenAI gender gap directly. Rather than launching a separate initiative, employers can incorporate AI upskilling into a process that is already mandated. Employers not subject to EEA requirements can adopt these measures voluntarily.

Practically, employers can integrate AI literacy into their EE plans by taking the following steps:

• Extend the workforce analysis to include AI literacy by capturing GenAI usage and training access by gender and other designated grounds.
• Frame low adoption or unequal training access as an employment barrier within the EE plan where the analysis supports this conclusion.
• Implement targeted training programmes linked to the EEA’s requirements for training and development to address identified barriers.
• Set measurable milestones by incorporating AI training targets into the EE plan’s year-on-year objectives.
• Monitor and report on progress as part of ongoing EE plan reporting.

The value of this approach lies in its simplicity: it takes an existing obligation and uses it to address a problem that might otherwise go unnoticed until the costs become difficult to reverse.

Conclusion

The gender gap in GenAI adoption is not merely a matter of individual preference – it carries real commercial, operational and compliance implications for South African businesses. Lower adoption rates among women can reduce productivity gains, entrench bias in AI systems and, for designated employers, undermine the transformation objectives they are required to pursue under the EEA.

By incorporating AI literacy into their barrier analyses and EE plans, businesses can turn a potential vulnerability into a strategic advantage – capitalising on the transformative potential of GenAI while fulfilling their transformation obligations under the EEA.