Women perceive more bias in hiring

In a survey of 4,417 people in 28 European countries, women, older people, non-white people, those on temporary contracts and those who identify as another gender reported a higher probability of experiencing negative discrimination in automated hiring processes. This indicates that built-in biases disproportionately affect groups that are already vulnerable.

In recent years, automated screening algorithms have become the first gateway for thousands of candidates in Europe. Companies that rely on AI systems to filter résumés promise efficiency, but practice has shown that these machines reproduce the prejudices they are loaded with. When a young professional in Madrid sees her application vanish without explanation, or when a temporary worker in Lisbon gets an automatic rejection, the problem stops being anecdotal and becomes structural.

The paper published in *Gender Work and Organization* addresses that gap with data from a survey of 4,417 people in 28 European countries, applying a feminist framework that makes it possible to cross gender, age, race and contract type. The authors asked respondents whether they had perceived discrimination in automated hiring processes and analyzed the answers with statistical models that control for sociodemographic variables. The results show that women, older people, non-white people, employees on temporary contracts and those who identify as another gender reported a significantly higher probability of experiencing negative discrimination from AI, compared with young, white men with permanent jobs.

The message is clear: it is not enough to buy the most modern tool. It has to be audited thoroughly and its decision criteria reconfigured. Incorporating datasets that reflect the real diversity of the workforce, setting up periodic bias reviews and opening the algorithms to outside review are essential steps. Creating feedback channels where employees can report automated decisions without fear of retaliation also makes it possible to detect and correct deviations before they become institutionalized.

Conclusions

• Hiring policies must review training data to keep it from reproducing gender and racial inequalities.
• Employers must include disparate impact audits before deploying screening tools.
• Incorporating feminist perspectives into the design improves trust and the perception of fairness among employees.