Female-dominated occupations are twice as likely to be exposed to generative AI as male-dominated roles, according to research by the International Labour Organisation. Yet women have about 22% lower odds of using GenAI tools than men, according to research by Harvard Business School.
The research was discussed during a recent virtual event hosted by South African law firm Cliffe Dekker Hofmeyr (CDH), which focused on SA's employment equity framework and the implications of GenAI adoption.
Safee-Naaz Siddiqi, senior associate for knowledge management at CDH, cited Deloitte research showing that only 61% of women who use GenAI feel their employer actively encourages it, compared with 83% of men. In addition, 49% of women say their company invests in GenAI training for them, compared with 79% of men – a 30 percentage-point gap.
"The group that faces the highest displacement risk is also the group least likely to be building the skills they need to pivot into the opportunities to replace those jobs," Siddiqi said. "This is not just an equality and inclusion conversation. It’s not a compliance issue. It’s the fact that half the talent pool – not just in your organisation, but the country – might not be building AI fluency, so that is shrinking the pipeline for succession by 2030. It’s not an HR problem, it’s a planning problem, it’s a board problem."
Siddiqi also referred to the World Economic Forum's Future of Jobs report, which projects that by 2030, 92 million jobs will be displaced globally by AI and automation, while 170 million new jobs will be created – a net gain of 78 million jobs.
"The reality is that 92 million jobs that exist today will no longer be relevant," she added. Asked who would be best positioned to move into new roles, she said candidates would need AI fluency, technological literacy and adaptability.
Fusion of two key aspects
Dr Nadeem Mahomed, director of employment law at CDH, said the research highlights the convergence of two previously separate issues: technology adoption and disruption on the one hand, and compliance with transformation and employment equity legislation on the other.
"Compliance is a legal obligation for employers, it is not something aspirational," said Mahomed, adding that the gender gap in GenAI adoption is not just a technology story but a potential employment barrier – "the very kind of barrier that designated employers are legally required to identify and address under the Employment Equity Act".
Rapid AI adoption
Retha Beerman, director for knowledge management at CDH, said AI adoption in the legal industry has surged. Citing Deloitte research, she noted that in 2024, 76% of respondents reported no AI adoption. "In 2026, only 2% reported no AI adoption, while 61% reported active deployment. That is a massive utilisation."
Beerman said many companies are adopting a crowdsourcing approach to AI in the workplace, partly because of the speed of adoption. Traditionally, technology was introduced with clear guidelines on usage, access and deployment. Today, however, AI is often "let loose" for employees to experiment with.
"AI is simply given to people, and they are told to go 'play with it' and find the efficiency within the process. We want them to explain that efficiency so we can extract it, replicate it and make others more efficient – in other words, crowdsourcing," she said.
"Those who use the technology right now are going to be the ones that step into the opportunity," she added.
In a report co-authored by Siddiqi and Mahomed and released in April 2026, CDH warned that the gender gap in GenAI adoption carries real commercial, operational and compliance risks for South African businesses. Lower adoption rates among women, the report found, could reduce productivity gains, entrench bias in AI systems and, for designated employers, undermine transformation objectives under the Employment Equity Act. However, the report also noted that businesses could turn this vulnerability into a strategic advantage by incorporating AI literacy into their barrier analyses and EE plans.
Not a confidence problem
Johan Steyn, AI expert and AIforBusiness.net founder, said women are bearing more of the disruption while benefiting less from the advantages.
"That is not a confidence problem to be lectured away – it is a structural one, driven by who gets encouraged, trained and given permission to experiment," he said. "Deloitte's finding that women are markedly less likely to be encouraged or trained in GenAI at work tells you exactly where the intervention needs to happen."
Steyn said AI fluency is fast becoming a core leadership competency, not a technical niche. "If half your talent pool is not building that fluency, you are not merely facing a diversity problem – you are quietly hollowing out your own succession pipeline for 2030 and beyond. In a South African context, where we are already fighting to build an inclusive skills base, allowing this gap to widen would compound an inequality we can least afford."
Steyn agreed in part that companies are adopting a crowdsourcing approach to AI, but cautioned that open access is not the same as equal access.
"A pure crowdsourcing approach rewards those who are already confident, already encouraged, already told it is acceptable to experiment – and that is disproportionately not women, not older workers, and in our context, not those from under-resourced backgrounds," he said. "Left unmanaged, it does not democratise capability; it amplifies the existing gap. The organisations getting this right pair the freedom to experiment with deliberate, structured enablement, so that empowerment reaches the whole workforce, not just the already-fluent."
Steyn stressed that the goal is to build capability across the entire workforce, rather than concentrating expertise among those already inclined to embrace AI.
Share
