Women are disproportionately represented in UK manufacturing roles exposed to generative AI, raising concerns that the technology could reinforce existing inequalities in hiring, pay and career progression, according to new research by appliance manufacturer Beko.
The Default Setting: Women, AI and the Future of Work in UK Manufacturing report found that an estimated 41.1% of women employed in UK manufacturing work in occupations exposed to generative AI, compared with 23.6% of men.
The research also found that women account for 27.8% of the UK manufacturing workforce but make up 63.4% of those in the occupations most exposed to generative AI. These roles are predominantly administrative positions, where women significantly outnumber men. Sales and customer service occupations also have significant exposure to AI, with women accounting for around half of the workforce in these roles.
The findings suggest that the impact of AI on manufacturing’s workforce will extend beyond production environments, with administrative and customer-facing roles potentially undergoing significant changes as generative AI becomes more widely embedded in business processes.
Concerns over AI bias
The Beko research also found a significant gender difference in perceptions of the risks associated with AI. Women were more likely than men to believe that AI bias could harm employee wellbeing (41.1% compared with 28.7%), pay and job opportunities (33.3% versus 22.9%) and diversity and inclusion (27.7% versus 17.1%).
Women were also less likely to trust AI-driven decisions, with 32.1% expressing trust compared with 42.9% of men. The differences remained even after accounting for factors including respondents’ roles, company size, age, AI use and level of expertise.
The report draws on a survey of 890 UK manufacturing professionals, analysis of Office for National Statistics employment data and International Labour Organization data on occupational exposure to AI, alongside a review of international evidence on AI and women’s careers.
While similar proportions of women and men said they had witnessed AI bias – 24.7% and 22.0% respectively – women were more likely to identify specific consequences. Some 28.3% of women identified flawed hiring decisions as a potential consequence, compared with 20.5% of men, while 16.7% of women highlighted the potential loss of talent through recruitment or promotion decisions, compared with 11.5% of men.
A lack of female representation in both the AI industry and manufacturing was also identified as a barrier to tackling AI bias by 24.4% of women, more than twice the proportion of men (10.2%).
The research also found concerns about AI-related pay inequality among younger women. Some 34.9% of women aged 22–29 identified pay inequality as a concern, compared with 39.0% of women in their 50s, despite the manufacturing pay gap being substantially smaller among the younger group.
Lack of visibility over AI policies
Despite growing awareness of AI’s potential impact on work, many employees remain unclear about how their employers are responding. Indeed, some 58% of respondents said they were unaware of any measures their company was taking to address AI bias, while around half of both men and women believed their employer was doing too little to prepare for AI. There was, however, broad agreement on the need for greater transparency, with 87% of both male and female respondents saying they wanted AI systems to be more transparent.
The report found that greater organisational preparedness was associated with higher levels of trust in AI-driven decisions. However, training and readiness did not reduce concerns about pay and job opportunities, suggesting that technical preparedness alone may not address wider workforce concerns. One female senior manager at a manufacturing SME described encountering potential gender bias while using AI to support recruitment.
“When sifting through CVs for a technical managerial position, AI seemed to prefer proposing male candidates,” she said. “When I double-checked AI’s recommendations for the position, I found a more suitable female candidate did exist among the CVs supplied.”
‘A significant opportunity’
Teresa Arbuckle, Beko UK & Ireland Regional Managing Director, said AI could improve manufacturing productivity while also creating opportunities to improve workforce representation.
“AI, used in partnership with human expertise and decision making, presents a significant opportunity to help manufacturing businesses become more efficient and productive,” she said.
“But this should not be at the cost of greater representation. We need to take everyone on this journey, regardless of gender, and we want to see more women enter manufacturing businesses and continue to progress into leadership and specialist roles.”
Responsible AI adoption
The report recommends that manufacturers incorporate workforce impacts into their AI governance and adoption strategies. Its recommendations include conducting risk assessments, audits and monitoring that consider the impact of AI on different employee groups; retaining human oversight and accountability over AI-supported decisions; and providing training and development opportunities for employees whose roles are affected by AI. It also recommends ensuring women are represented in decision-making around AI through leadership teams and employee networks.
Beko said it had reviewed the recommendations and was using the findings to strengthen its existing responsible AI and people policies. Arbuckle said the company was also focused on ensuring that employees affected by technological change were able to benefit from new opportunities.
“As the world of work changes, and we explore how AI can support our business and customers, we are continuing to strengthen the policies, training and support available to our people, ensuring those most affected by change are considered and are able to benefit from the opportunities it creates,” she said.
“It is essential that the progress we have made on women’s representation is embedded into the new processes and technologies that we adopt.”
For more articles like this, visit our Data & AI channel
