Gender bias in hiring processes assisted by artificial intelligence is often underestimated, according to a study by Belgium's Institute for the Equality of Women and Men.
The research, conducted by the University of Liège and Hasselt University, found that three-quarters of recruiters use artificial intelligence at some stage of the recruitment process.
Although AI is perceived as more objective, the analysis indicates it can reproduce stereotypes and cause discrimination, potentially exacerbating existing inequalities.
A minority of recruiters – around 12 to 17% – acknowledged observing biased outcomes from AI, especially linked to social media algorithms, age, and ethnicity. However, gender bias is rarely cited spontaneously.
The Institute has called for stronger legal frameworks, noting that only 20% of organisations have implemented measures aligning with AI legislation. The European AI Act has identified recruitment-related systems as "high risk," requiring specific safeguards.
The Institute also recommends assessing whether Belgium’s anti-discrimination laws are suited to AI-related challenges in recruitment and workplace practices.
Later this year, the organisation plans to release a brochure offering employers guidance on using AI responsibly in the workplace.
The study involved over 400 recruitment professionals.
