The recruiting departments of large companies are hiring artificial intelligence.
The computer even knows how to find out if the applicant is lying in the interview.
All right, we have arrived. The recruiting departments of large companies are hiring artificial intelligence (AI is the term used in everyday life) to find employees. Can't cope on their own? Or do computers pick people better? It seemed to me that such a delicate matter as the search for employees cannot be entrusted to a soulless machine - here human relations are important. But now this area has already been "digitalized" … Recently I met the guys who are developing such a system in Russia.
“Every year it’s more and more difficult to find suitable employees,” explains Evgeniya Dvorskaya, co-founder and CEO of such a service Sever. AI. - There are several reasons. The main one is the demographic pit of the 90s. There are very few people between the ages of 25 and 35 who have not yet found a job. And this is not an excuse for lazy recruiters, but Rosstat data.
It would seem that low competition should make life easier for potential employees. But - paradoxically - finding an interesting, suitable job is not easy. On the one hand, thanks to services like Headhunter, SuperJob or Avito. Job, there are much more opportunities for finding a lifelong business. These are not advertisements that you had to hang on poles.
But the number of companies that post jobs online has also increased.
And one more serious problem for HR specialists - there are more resumes of potential candidates. Previously, it was customary to write one, but now, if a person believes that he could work in several areas, he writes a separate resume for each employer, emphasizing the experience that he has to obtain this particular position.
As a result, recruiters bury themselves in a heap of profiles. And they miss the right person. On a sunny morning, a person seems like a suitable candidate, and in the evening he can be severely rejected.
Promotional video:
After the first selection of suitable resumes, the recruiter calls the candidates. The most suitable ones are called for an interview or skype interview. A half-hour conversation turns into a day of hours if you need to communicate with 50 candidates!
In general, employers spend a huge amount of money and time looking for the right people. At the same time, sometimes without an obvious result.
The robot has no bad days
And now, instead of a person, artificial intelligence begins to rule the ball here …
- Yes. But we're talking about really smart systems. A robot that offers a designer job to anyone who knows how to use Photoshop is very far from artificial intelligence, explains Dvorskaya.
Artificial intelligence is different. He learns from the experience of real recruiters - how they evaluate resumes, what they pay attention to in interviews. And then he himself looks for suitable candidates and does it instantly. AI has capabilities that we do not have: for example, it can quickly figure out that several interesting resumes belong to the same person. The robot has no good or bad days, it is always objective and uses the best skills of its human teachers. He sets priorities himself and calls motivated people first. And most importantly, all this takes a few minutes: he can conduct thousands of conversations and video interviews simultaneously.
Understands even emotions
Another plus is the objectivity of the candidate's assessment. The robot quantifies the resume and video interviews on the parameters that are important specifically for this future employee.
Wait, it's one thing when a robot crawls the websites of recruiting companies and looks for the right resume, another thing when it - as you said? - ringing up applicants? It turns out that I have to interview a computer?
- Yes, it has already come to that, - says Dvorskaya. - And during the interview, Sever. AI analyzes facial expressions and voice of a person by 128 parameters. He understands the logic of what the future employee says just like the character of Tim Roth from the TV series "Lie to Me"! And marks it out. According to the assessment given by the AI, the employer can understand whether to hire a person or not.
Well, you can't find a scientist to work in a laboratory, a teacher in a university …
- Of course, we are talking about finding people for mass, linear positions: cashiers, salespeople, call center operators. But it is quite possible in this way to hire more qualified employees - sales managers, bank tellers.
I know that in many companies, before inviting a job seeker for an interview, an HR officer studies his social pages. Does AI do this?
- Yes, we are testing this possibility. It is clear that when looking for a seller, it is hardly worth analyzing his entries on VKontakte. But if a person is being looked for for an important position, for the place of a leader, social networks provide useful information.
This is not a lie detector
If the AI is able to analyze the voice and facial expressions of a person, it means he understands when he is lying … We have not been caught for lies yet …
- Of course, AI is not a lie detector. Our main task is to understand whether a candidate is suitable or not, find those who are needed, and conduct the first structured interview. Although a robot is indeed much more attentive than a human recruiter, it treats details - phrases on a resume, a job seeker's behavior in a video interview, his emotions and reactions to specific questions. The candidate has much less chances to deceive the computer - he will understand this by indirect signs.
Lying on your resume and in the interview is a really big issue. A person does not always do this consciously - it just seems to him that it is better to embellish something, to make himself more significant in something. By the way, we conducted such a study. Most often, candidates lie in their resume (45%), less often in the interview itself (32%). Usually, candidates for line positions (64%) make lies, less often - managers (31%).
And in what lies usually?
- More often they deceive when describing their experience - (79%), try to tell that in their previous job they held an influential position and received a higher salary than they actually did - (46%). But, of course, the reason for leaving the previous company is the incompetence of the local leadership - (43%).
But to be honest, employers also sometimes cheat candidates. When it comes to the exact amount of income or growth prospects (26%), the need for travel (25%) or a set of responsibilities (23%).
Of course, if a person sets out to deceive the employer, then he, most likely, will be able to do it - neither the recruiter nor the robot can get into the brain. People even lie detectors cheat.
Twice as efficient as humans
You know, I am a person of conservative views, and I do not really like it when a soulless computer will select me for work. How much can you trust a machine when it comes to human relationships?
“Our tests show that a machine is 1.5-2 times more efficient than a person in finding the right people. Recruiters considered 9 out of 10 candidates found by AI to be suitable. At a time when usually a recruiter has a result of 6-7 suitable candidates out of 10 is considered good.
And then, the final decision on hiring remains with the person. Maybe in the future, when people begin to trust robots more, for some positions the decision can be made completely automatically. HR will be able to tackle more important employee issues.
What about personal data protection? Are you warning you about using technology like this?
- We ask the candidates for permission to process the information that will be obtained during the video interview. We do not share this information with third parties in any way.
In what form does the machine give an opinion?
- There are two ratings on whether the resume is suitable and how the results of the video interview correspond to the tasks for which the applicant is planned to be accepted.
Browser comment
The machine will calculate your career
Large companies have been using intelligent recruiting systems for several years. The first signs were chat bots (programs that are trained to ask questions on a specific topic and "understand" the answers received). There are several retail chains in Russia that conduct the initial selection of employees using such chat bots. The use of AI that can analyze the structure of the resume and the results of video interviews is a feature of recent times.
Setting up an AI to do this is not easy. The recruiter uses specific filters to find the right resume. For example, if his company recruits for a vacancy with a salary of 35 thousand rubles, then the resume in which the applicant writes that the expected salary is 37 thousand will be excluded from the results. But the AI of such a candidate will offer the recruiter if other parameters match the request. Why not bargain?
The massive use of AI and neural networks for recruiting is a matter of the near future, get ready. Well, the next step will be AI consultations to the bosses - which employee should be promoted up the career ladder, who should be enrolled in the talent pool, and who should be eliminated as ineffective. The machine will learn to analyze not only the current skills of the employee, but also how promising he is for the company. The computer will rank the staff, calculate the "correct" salary and career path. And this is no longer a fantasy. Alas.
ALEXANDER MILKUS