Australia’s discussion about artificial intelligence and the workforce is frequently framed around productivity: how can AI help employees complete tasks more efficiently, address skills shortages and support economic growth? Two recent Australian articles argue, however, that the productivity benefits of AI will be constrained unless women can fully participate in its adoption. Sally Browner, writing in Women in Digital, focuses particularly on women aged over 55, while gender equality advocate Kit McMahon examines the broader structural barriers affecting women’s participation in an AI-enabled economy.
Both argue that the challenge is not simply to teach more people how to use AI. Adoption is also shaped by trust, workplace culture, unpaid caring responsibilities, occupational segregation, experiences of discrimination and whether people believe the technology, and the systems governing it, will operate in their interests.
Evidence of a gender gap in AI use is becoming increasingly clear. A Harvard Business School synthesis, cited by Browner, of 76 sources covering more than 100 countries found that gender gaps in generative AI use were widespread and persistent, including among people working in the same occupations and organisations. Browner argues that age can compound this divide: older women may have lower confidence in their capacity to use new technologies, less trust that AI providers will protect their data, and less time available for learning because of responsibilities for children, grandchildren or ageing parents. Yet older women are not a marginal part of the workforce. Citing Australian Bureau of Statistics data, Browner notes that the labour force participation rate of women aged 55–59 increased from 52.1 per cent in 2004 to 73.5 per cent in 2024. Failing to include these experienced workers in AI capability development therefore represents a significant loss of knowledge, judgement and potential productivity.
McMahon similarly challenges the assumption that women’s lower use of AI can be resolved through generic digital literacy programs. Research cited in her article suggests that improving technical knowledge alone may not increase women’s adoption when concerns about privacy, bias, job security and the social effects of AI remain unaddressed. Women’s caution can also be a rational response to technologies that have reproduced discriminatory patterns, enabled gendered online abuse or produced biased recruitment and assessment outcomes. McMahon also highlights evidence that women who disclose using AI can be judged more harshly than men producing work of the same quality. At the same time, occupational segregation means women are disproportionately concentrated in roles likely to experience significant AI-enabled augmentation or redesign. Training is essential, but it cannot by itself compensate for unsafe technologies, biased evaluation practices or workplaces that place responsibility for adapting entirely on individual employees.
For VET educators, these arguments have important implications for the design of AI learning. Capability development should be connected to authentic workplace tasks and be designed for learners who may be time-poor or beginning with different levels of confidence. Programs will also be more inclusive when women, including older women, women with disability, First Nations women, regional learners and those with caring responsibilities, are involved in determining how training is delivered and what support is needed.
VET leaders also need to examine AI adoption within their organisations. Meaningful approaches to understand AI use could include protected learning time, approved and secure tools, practical mentoring, as well as psychologically safe opportunities to experiment. Participation and outcomes should be examined by gender, age, employment status and other relevant characteristics, while women should be represented in decisions about AI governance, procurement, curriculum and workforce redesign.
VET has a dual role in the AI-transition: preparing learners for changing workplaces and modelling fair adoption within our own workforce. As Browner and McMahon argue – equity is not an additional consideration to be addressed after productivity gains have been achieved; it is one of the key conditions required for those gains to be realised across the whole economy.
