Tech recruiting firm uses face-scanning algorithm to make hiring decisions
Marie Donlon | October 28, 2019Some U.S. companies are reportedly using artificial intelligence (AI)-based face scanning algorithms to make hiring decisions.
Tech recruiting firm HireVue is helping industry giants — particularly those in the hospitality and finance markets — to make hiring decisions based on a candidate’s facial features. HireVue analyzes a candidate’s video interview conducted using either the camera on the candidate’s smartphone or other device. Once the 30 minute interview is completed, the company applies its algorithms to the video, assessing candidates on the basis of their facial expressions, the sound of their voice and the words they select in the course of the interview.
The algorithm reportedly analyzes the video content, picking up on factors such as the candidate’s enthusiasm, eye contact and roughly 500,000 other data points that are largely undetectable to human hiring professionals, to make determinations about the candidate’s overall productivity and thus employability. Once assessed, the candidate is given an “employability score” that is then measured against other candidates’ employability scores.
Among the U.S. employers already using the HireVue services are Goldman Sachs, Unilever and Hilton, to name just a few. So far, these employers suggest that the AI-driven system has reduced the amount of money and time they spend making hiring decisions.
AI-driven hiring decisions are reportedly becoming so commonplace that U.S. colleges and universities are preparing students for the inevitability of such automated decision making by emphasizing how they should comport themselves during video interviews that will be assessed by AI. According to HireVue, appearance alone accounts for 29% of an applicant’s employability score while word selection and the sound of the candidate’s voice make up the rest.
Despite HireVue’s claims that its algorithms undergo regular bias-testing, critics of the system suggest there is danger in making hiring determinations based on outward appearance, suggesting that it encourages discrimination. For instance, a candidate who is nervous, who doesn’t speak English as his or her native language, or who challenges existing standards of attractiveness might receive poor employability scores and, consequently, be overlooked by the hiring algorithm.
For better or for worse, these companies that use this technology will get the candidates they deserve.
<...a candidate who is nervous, who doesn’t speak English as his or her native language, or who challenges existing standards of attractiveness might receive poor employability scores and, consequently, be overlooked by the hiring algorithm...>
It needs an <...algorithm...> to do that?
This approach to hiring decisions seems very likely to enable corporate HR and executive decision-makers to insulate themselves from accusations of using illegally biased, in-house "created", corporate owned AI intellectual property" in making such decisions. Using an externally developed, "off-the-shelf" AI-based HR App would certainly muddy the waters in any liability decisions for use of illegal biases in hiring practices. The entire, broad issue of liability, either on the AI App developer or the App using enterprise must be clearly defined and documented.