Home / Technology / Belgian Study Flags AI's Hidden Gender Bias
Belgian Study Flags AI's Hidden Gender Bias
4 Feb
Summary
- Most Belgian recruiters use AI in hiring, but few see bias.
- AI can perpetuate stereotypes, worsening inequality.
- Belgium needs stronger legal protections for AI tools.
A new study conducted in Belgium indicates that gender bias within AI-powered recruitment tools is frequently underestimated by employers.
The research, a collaboration between the Institute for the Equality of Women and Men and two universities, surveyed over 400 recruitment professionals. It found that approximately three out of four organizations now utilize AI in some capacity during hiring processes, such as screening applications.
Despite a common perception among recruiters that AI is fairer than human judgment, the study revealed its potential to replicate societal stereotypes and foster discrimination. Only about 12-17% of recruiters reported observing biased outcomes from AI, with most noting issues related to social media algorithms, age, and ethnicity rather than gender.
The Institute has issued a warning, asserting that Belgium requires more robust legal safeguards for AI technologies. Currently, only a fifth of organizations have policies aligned with existing AI legislation. The European Union's AI Act categorizes recruitment tools as high-risk, mandating stringent checks and bias mitigation measures.
Furthermore, the Institute recommends a review of Belgium's anti-discrimination laws to ensure their effectiveness for AI applications in employment. A guide to responsible AI use for employers is slated for publication later this year.
