Friday, 18 august 2017 | Redacción CEU
The Human Resources professionals increasingly trust on the algorithms to carry out their applicant selection. The information to process is as extensive as overwhelming; this procedure automation to choose candidates sometimes supposes a life jacket so the recruiters do not drown. Meanwhile, the artificial intelligence systems begin to occupy greater responsibility positions like conducting the screening interview to the applicants. Are robots the most suitable 'subjects' to find the talent?
You do not like talking to machines. When you phone a company and a robotic voice answers, you get angry –do not get worried, it happens to all of us–. Whether you like it or not, the automation and the bots use are increasingly integrated in our society. The good news is that these systems are exponentially more intelligent as time goes by, because you have to recognize that what really bothers you about that call is that only an operator is able to understand you. Are you still annoyed? You can always hang up... But if you must talk with a robot for getting a job, the resignation is not as easy.
Like the Professor Manuel A. Alonso Coto explains in his masterclass e-Recruitment: digital business optimization, the reach of this artificial intelligence has advanced a lot in recent years. He takes the example of the robot Jill Watson which posed as a professor at The Georgia Tech University and obtained incredible results, so much so that her students did not realize during the online course that they were actually talking to a robot. Watson was a teacher who answered questions, gave explanations and raised debates under her digital disguise. Her performance was so good that it was even nominated by some student as the best university teacher.
The success of these artificial intelligence systems lies in the natural language understanding improvement, but if these technologies want to resemble the flesh and bone professionals, they still have to overcome certain obstacles. At least, that is the conclusion drawn from another experiment carried out by the developers of the robot Torobo-Kun which they subjected to the Japan university entrance exam. Although this surpassed the students average, it was not able to obtain place in the Tokyo University –considered the best one of the country–.
According to Noriko Arai, Specialist in Artificial Intelligence and the Project Searcher, all modern artificial intelligence are statistical machines that collect information from the human world through different sources like textbooks, Wikipedia, Facebook or Twitter. Actually, they do not understand anything, they only say the right thing for statistics and select the keywords according to the question clues. Even so, they outperform most students. The challenge of this sector is focused on teaching these systems to germinate their creativity and to reflect and this can only be done through the mathematical language development.
Nowadays, the vast majority of companies employ systems that use algorithms to scrutinize and select the resumés of the most appropriate candidates for the offered positions. The work of the recruiters would sometimes consist of a titanic work without these tools and with the amount of information that can be generated. The tasks entrusted to the robots acquire greater weight like making the first screening of candidates or carrying out the own interview.
The British startup Cognisess is dedicated to perfecting this technology. The InterContinental Hotel Group is one of the first customers in using its robotic screening to find the suitable staff. The company has outlined the ideal manager profile and the system follows it to carry out the first selection tests. According to its developers, the program is able to detect facial expressions and basic emotions through the data extraction and even determine if the words of the candidates contradict what they feel.
There are 'intelligent' machines capable of conducting a complete job interview. They can even play music to the aspirants in order to calm them down, like the robot MatIlda. Rajiv Khosla, Director of the Research Center for Informatics, Communication and Industrial Innovation at La Trobe University and the creator of the device, argues that what makes it different from his human counterparts is that decision-making in the recruitment process is free of whim or prejudice.
Algorithms reduce prejudices in hiring. At least, that is the premise advocated premise by many of the developers of these systems. However, we cannot overlook that the people who program these robots also are subjective. It is very difficult to avoid partiality in their design. Of course, humans are not equable in recruitment either –that is the paradox–.
Until now, the 'smart' programs are not capable of perceiving sympathy and kindness. In a human interview, being liked can lead an unsuitable candidate to get a job, just the opposite can also happen. However, these factors can be decisive in the choice of a professional who works work face-to-face with customers.
A factor to take into account in the selection process is the own candidate's experience. Companies should take care of the first contact with potential employees. If robots are the first impression taken by the applicants, these 'smart' devices will have to know how to arouse the interest of the best ones so they do not go with the competition firms.
Interviews are a kind of two-way communication. The interviewees also draw their own conclusions in the interaction with the recruiter: if the job fits with their aspirations, if they have projection, if they like the image of the company,... The dialogue with a smart system is still far from looking like a conversation with a human recruiter. If these selection methods want to attract the best professionals, they should also adapt to the applicants interests.
When robots become creative, they reason and reflect, people will no longer be part of the equation. Until then, the last word will always come from a human mouth.