Pages

2018-05-28

Luddites lost, hopelessly: Artificial Intelligence now helps recruit staff and detects depression

More and more German companies rely on artificial intelligence when it comes to personnel selection. Insurer Talanx has stated that it conducts executive search with software that creates within minutes extensive personality analyses of applicants based on language tests. According to Talanx, the software produces a 90 percent approximation of the results obtained by psychologists after days of work at assessment centers. Talanx relegates assessment to an algorithm for good reason: a large part of its management will retire by 2025. Other companies using the same software, Precire, include Frankfurt Airport’s operator Fraport and Ranstadt HR agency. The algorithm test of Precire was created by an Aachen startup of the same name. The developer currently works on integrating a medical speech analysis program for early detection of depression. To date, Precire is able to reveal first signs of depression at a very early stage.

Artificial intelligence development is considered a multi-billion-dollar market in Europe, and the situation in the U.S. is very similar. AI technologies could soon make impact on other sectors, but platforms such as Precire are both highly promising and alarming at the same time, as cheap and increasingly accurate technology is bound to spread quickly.

Alas, practical concerns reach far beyond EU’s GDPR: who will ensure that this easily recorded deep psychological language analysis will be used only with valid consent and for certain purposes but not for others, and not by potential anonymous actors? More immediately, who will incur the risk of hiring an individual with symptoms, or even likelihood of future onset of depression? What else besides depression will be revealed by additional modules and additions to such a platform? What are the consequences for Fourth and Fifth Amendment rights? Do individuals retain a reasonable expectation of any privacy if physical characteristics such as voice and images are increasingly available from plenty of uncontrollable recordings, such as millions of CCTV cameras, public and private webcams, and voice-operated AI systems? Will banks start to rely for credit decisions also on customer profiles based on voice commands given to automated navigation systems? How about individuals considering relationships with new acquaintances? Has anyone ever seen a genie retreat into its bottle?

Multiple elections in Western countries since 2016 were about “the forgotten men and women,” the losers of globalization and victims of automation. It is safe to say that their numbers are bound to grow while neo-Luddite resistance may destabilize purportedly open societies from inside more than migration has to date. As the potential for abuse by further dilution of traditional concepts such as “informed consent” and “reasonable expectation of privacy” proliferates, will resistance to technology, even at great sacrifices of convenience and price, remain even theoretically possible?