Digital MarketingAI better rates candidates who do job interviews with background books
According to research by BR, in the selection processes the AI gives a better score to candidates who face video interviews with book funds.
The artificial intelligence (AI) tools used in the selection processes (carried out mostly in a virtual way in these times) are often not as neutral as might be assumed. In order to change the AI’s judgment with respect to a candidate, it is enough, in fact, to choose a certain fund and not another.
We are facing a statement that is inevitably controversial if we consider that companies that rely on AI for their selection processes argue that Artificial intelligence is not guided by tastes of a personal nature, it does not obey impulses when making decisions and is in this sense absolutely neutral.
However, according to the Bavarian public television and radio company (Germany), using the AI tools of the Munich-based startup, it seems that artificial intelligence is far from being neutral in selection processes and he throws completely different results on the table when he has in front of him candidates equipped with glasses or headscarves.
AI judges candidates differently based on their plugins and also the background images used in interviews
In the context of the research undertaken by BR an actress recorded a short video as a nomination that was later classified by the Retorium AI based on the so-called “Big Five” model, which analyzes the personality of the individual guided by five different dimensions.
The actress then read the same text with the same expressions and the same tonality on several occasions. The only thing that actually changed is that on one occasion the candidate wore glasses and on another she wore a scarf covering her hair. On both occasions the results from the “Big Five” analysis were completely divergent.
Further attempts with modifications in the form of wigs, ponytails and other head ornaments also resulted in substantial changes. in the analysis produced by the «Big Five» model.
In light of the results of BR’s research, Retorio points out that its artificial intelligence also evaluates the candidate’s outward appearance, as it would also be done in any usual job interview.
Nevertheless, Retorio’s algorithms only evaluate the impact and impression that one person makes on others, excluding potentially biasing factors such as age, gender or race.
In subsequent tests BR experimented with different background images (a framed painting, for example) in the videos of the candidates and the results were also at the mercy of many modifications.
If the candidate does, for example, the job interview with a bookstore as a background image, the analysis of Retorio’s algorithms changes and does so in favor of the interviewee. And it also does when video lighting is improved.
Deviations committed on behalf of AI are not always systematic and may vary from person to person. “The fundamental problem with facial recognition through ‘machine learning’ is that we never know exactly based on what patterns the algorithms react in an image,” says Katharina Zweig, professor of computer science, in statements to BR.
In this way, certain regulations would be necessary in the field of machine learning and artificial intelligence, Zweig asked.
Don’t miss anything from MarketingDirecto.com and join our Telegram
Source: Marketing Directo by www.marketingdirecto.com.
*The article has been translated based on the content of Marketing Directo by www.marketingdirecto.com. If there is any problem regarding the content, copyright, please leave a report below the article. We will try to process as quickly as possible to protect the rights of the author. Thank you very much!
*We just want readers to access information more quickly and easily with other multilingual content, instead of information only available in a certain language.
*We always respect the copyright of the content of the author and always include the original link of the source article.If the author disagrees, just leave the report below the article, the article will be edited or deleted at the request of the author. Thanks very much! Best regards!