Utrecht University: Artificial intelligence assesses applicants more fairly in terms of motivation
Naber and his colleagues conducted a NWO-subsidised study to assess the motivation of job applicants. In collaboration with the company Neurolytics, he tested a computer model that scans complex behaviour and emotions in the video recordings of the applicants.
online solliciteren
Computer model is fairer
In Naber’s experiment, 154 test subjects were tasked with practising an online application for a job at a company. Afterwards, they were asked how motivated they were during the application. Naber: “The computer model detected that less motivated applicants behaved differently from the highly motivated applicants during the interview. Human assessors did not see this and often labelled motivated applicants as ‘demotivated’. Some assessors even wrongly judged that men were more motivated than women.” Naber claims that a computer model, provided it is trained on the basis of feedback from the applicants, assesses people more fairly than recruiters and human resources managers.
Prejudices are a persistent problem when hiring and promoting employees.
Subtle muscle changes
Among other things, Naber’s computer model showed that demotivated applicants often contracted the muscles in their chin and lips for a relatively long time. “Rather as if they were stiffening their jaw.” The combination of different subtle muscle changes in the face was particularly informative. “Each individual expresses behaviour in a slightly different way, but the computer model also learned this. The human assessors did not know which facial information to look out for. This is very difficult to learn, even for people who regularly come face to face with applicants. However, artificial intelligence makes this possible.”
Discrimination in the workplace
The Utrecht-based researcher believes that this new technology can be of great value during the assessment process. “We have known for some time that discrimination exists in the workplace: prejudices are a persistent problem when hiring and promoting employees. Artificial intelligence can now offer a solution.”