In a recent study published in Nature, researchers at the London School of Economics (LSE) used artificial intelligence to analyze the behavior of recruiters on job sites. What their AI discovered is that the recruitment system often hides unconscious biases.

The LSE’s AI tool, which uses supervised machine learning algorithms, was applied to a Swiss public service job board. The research used data from 452.729 searches from 43.352 recruiters, 17,4 million profiles and 3,4 million CV views. The researchers then looked at how much time recruiters spent reviewing CVs, and when they decided to interview a candidate.

Candidates from immigrant and minority backgrounds are 19 percent less likely to be contacted by hiring managers, the study found. And these biases in recruiting also cut across gender lines. Women in male-dominated industries are 7 percent less likely to be called for interviews based on their resumes, and the same is true for men applying for jobs in a female-dominated field.

Another AI finding: The level of discrimination varies depending on the time of day. Hiring managers tend to spend less time reviewing resumes before lunch and near the end of the workday. At these times, immigrant and minority candidates are 20 percent more likely to be screened than people from the majority population with the same qualifications.

According to Dr. Hangartner, one of the authors of the study: “The results suggest that unconscious biases, such as stereotypes about minorities, have a greater impact when recruiters are tired and end up falling back on intuitive decision-making.”

To solve the problem, the researchers suggest that recruitment websites place less emphasis on candidates' names and nationalities, placing this information below education, qualifications and experience.

Through which channels you reach those people, classic and out of the box. The Next Web

Image: Fauxels / Pexels