Artificial intelligence could deepen gender inequality in the workplace unless women play a far greater role in shaping the technology, according to new research from the Women and Work All-Party Parliamentary Group (APPG).
The report, which draws on evidence gathered during a series of industry roundtables between 2024 and 2025, warns that AI systems trained on historically biased data could replicate and even amplify existing discrimination in areas such as recruitment, career progression and performance evaluation.
Researchers argue that without more representative datasets, stronger oversight and greater diversity among the people designing and deploying AI systems, the technology risks embedding workplace inequalities at scale just as businesses increasingly adopt automation and algorithmic decision-making.
The findings highlight several real-world examples where algorithmic systems have demonstrated bias. One case involved the withdrawal of an AI recruitment tool developed by Amazon after it was found to favour male candidates over female applicants. Concerns have also been raised about the visibility of women’s professional content on platforms such as LinkedIn, where algorithmic ranking has reportedly reduced the reach of posts written by women compared with those authored by men.
More broadly, experts say large language models and other AI systems frequently learn patterns from historical data that reflect longstanding gender imbalances in employment and pay. If those patterns are not corrected during development, the systems can unintentionally reinforce them when used in real-world decision making.
Support authors and subscribe to content
This is premium stuff. Subscribe to read the entire article.
