Artificial intelligence could intensify gender inequality in the workplace unless women have a stronger role in shaping the technology, according to a new report by the Women and Work All-Party Parliamentary Group (APPG). The research, based on evidence from industry roundtables conducted between 2024 and 2025, highlights risks that AI systems trained on biased historical data may replicate and even amplify discrimination in recruitment, career progression, and performance evaluation.
The report cites real-world examples of AI bias, including an Amazon recruitment tool withdrawn after favouring male candidates and algorithmic rankings on professional networks such as LinkedIn, which reportedly reduce the visibility of women’s posts. Large language models and other AI systems often learn from historical patterns of employment and pay, reflecting longstanding gender imbalances that can be reinforced if not corrected.
The study warns that women face a dual risk in the expanding AI-driven economy. They are underrepresented in technology leadership and development, yet overrepresented in roles most vulnerable to automation. Administrative, education, healthcare, and social care positions, sectors dominated by female workers, have been among the first affected by AI-driven automation. Experts say that as AI adoption grows, displacement risks could increase unless women acquire the necessary digital and technical skills.
Karren Brady, co-chair of the Women and Work APPG, said the acceleration of AI is reshaping labour markets at a time when gender inequality persists. “The enduring gender pay gap and the continued lack of parity within the technology sector make clear that meaningful progress remains unfinished,” she said, stressing that urgent action is required.
Industry leaders contributing to the report emphasised the importance of the data used to train AI. Linda Benjamin, vice president at AND Digital, said, “AI is shaped by the data it’s built on, the questions it’s asked, and the people who design it. When historical data reflects gender imbalances or systemic bias, AI can learn and replicate those patterns.” She argued that addressing these inequalities requires representative datasets, rigorous auditing, and greater female participation in AI and digital careers.
The report also highlighted structural barriers that limit women’s opportunities, including restricted access to reskilling, high childcare costs, and workplace structures that hinder movement into technical roles. Older women face particular vulnerability, as training programmes often exclude workers over 55, increasing the risk of redundancy.
Concerns were also raised about AI-driven productivity monitoring tools, which, without safeguards, could create punitive working environments. Charlotte Wilson, head of enterprise business UK and Ireland at Check Point Software Technologies, noted that AI offers benefits, such as early detection in healthcare, but requires careful oversight to prevent bias in workplace decisions.
The APPG’s ongoing research aims to develop policy measures to ensure women are not left behind as digital transformation continues. Recommendations include expanded digital training, improved childcare support for entrepreneurs, and stronger safeguards against algorithmic bias. Advocates argue that empowering women in AI development is essential for equality and economic competitiveness.
