When will AI replace engineers?

Artificial intelligence in recruiting: does it make recruitment consultants superfluous?

The digital wave is sweeping through all industries - including HR. Artificial intelligence evaluates applicants' résumés, analyzes job interviews and gives recommendations as to which candidates are best to hire and which are not. Can HR consultants surf the wave of digitization or are they being washed away by it?

Artificial intelligence can relieve HR departments of a lot of work when it comes to selecting suitable candidates for the first time. But that is only one possibility of robot recruiting.

Photo: panthermedia.net/AndreyPopov

How do companies use AI in recruiting?

"Some work steps can be automated very well", Nikola Hänsel admits frankly. Hänsel is a managing partner at Müller und Partner in Düsseldorf. The personnel consultancy has been on the market since 1990 and primarily places IT specialists. Using algorithms, you can quickly and automatically narrow down a large number of applicants, says Hansel. It's already happening today. If an applicant does not meet the so-called must-have requirements - a developer without Java knowledge, for example, or a sales representative without the willingness to travel - then he is thrown out without a person having even glanced at his application. "I assume that these search processes will be automated to a much greater extent in the future than they are now," said Hansel. She also observed that many companies used to consult psychologists who checked applicants for special qualifications or leadership qualities during the interview. Today that rarely happens. Online assessments have replaced people because they work faster and more cheaply.

This is how much artificial intelligence experts earn

Unilever is already one step further. According to its own statements, the Dutch-British consumer goods multinational saves around 100,000 manpower hours per year and thus hundreds of thousands of British pounds through the use of artificial intelligence in recruiting. Every applicant at the ice cream and shampoo manufacturer is guided through the application process by machines - and not by people. First, the applicant plays a few little games on the screen. Characteristics such as logical thinking or willingness to take risks are tested in this way. This is followed by a 30-minute video interview. The analysis software checks the candidate's facial expression, whether he or she smiles frequently or looks grim, evaluates body language and choice of words. If everything is to the satisfaction of the bot, the applicant goes one round. Only then do you meet real people of flesh and blood who finally raise or lower their thumbs.

In this way, Unilever filters 30,000 new employees out of a total of almost two million applicants each year. For this purpose, Unilever teamed up with the US startup Pymetrics from New York, which is building the AI ​​software. Pymetrics has raised over $ 56 million in seven funding rounds since 2012. New Yorkers are part of the growing number of HR startups that can use their algorithms to enrich - or destroy - the headhunter's business. In Germany's HR departments, too, AI tools have long been in the toolbox. Anyone surfing the careers site at BASF is greeted by a chatbot called Anilina, who answers questions about looking for a job around the clock. He was fed with the most common questions that Basf HR staff had been confronted with in five years of live chats. If Anilina cannot answer a question, a human employee takes over. The chatbot learns continuously and saves new questions so that it can answer them the next time. In the US, Basf works with the Utah-based AI company HireVue. HireVue's algorithms evaluate how candidates have done in video interviews, online games or programming tests and make a recommendation. In Germany, however, Basf does not use this software, as the company emphasizes to INGENIEUR.de.

Artificial intelligence in recruiting: what are the advantages?

Unilever and Basf are not the only companies that promise advantages in recruiting through the use of digital tools. The US job exchange Career Builder, for example, wants to reduce the recruiting time of its customers by 30% through artificial intelligence. 200 data scientists evaluated two billion data points over the course of a year. The ad placement is now a lean four-step process and completed in five minutes - thanks to the AI.

Data Scientist: What does the professional data evaluator do?

Many HR activities have shifted to the large business platforms Xing and LinkedIn. Both are making use of ever larger databases that they make available to their customers. The LinkedIn systems compare the user profiles with a job advertisement and then present the employer with a hit list with the best candidates. However, the service is not a substitute for traditional recruiters, as John Jersin, Vice President at Linkedin, pointed out as a precaution. The German counterpart Xing offers similar solutions. Candidates are shown job offers or sent to their mailbox by newsletter, which presumably fit like a pot on top - without any assistance from recruitment consultants. The search for personnel has changed completely for her in recent years, says Nikola Hänsel. In the past it was often enough to place a vile job advertisement, today it is mainly active on Xing and LinkedIn.

The greatest advantages of digital tools and artificial intelligence in recruiting at a glance. You…

  • save manpower hours
  • reduce costs
  • allow feedback to be given to each individual applicant
  • can process a larger number of applicants
  • can improve the matching

Artificial intelligence in recruiting: what are the disadvantages?

A year and a half ago, Amazon stopped the experiment. The online giant's newly developed recruiting system wasn't working as it should. Artificial intelligence gave applicants star ratings from one to five. The scheme was similar to the product reviews on Amazon. The ultimate goal was to use the AI ​​to filter out the best candidates for a job from all applications in a lightning-fast and super-simple manner. But the artificial intelligence was not impartial, it was explicitly misogynistic. Female résumés were punished because the system internalized the typical patterns in the IT industry and learned that the successful applicants so far had mostly been men. At the end of 2018, Amazon stamped the project.

The much-voiced hope that artificial intelligence in recruiting could remove prejudices from application processes has since given way to considerable doubts. But that's not the only problem with artificial recruiters. In an experiment, scientists from the Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) recently set up artificial chatbots against human interview partners. They wanted to know who the participants trusted more. The answers were identical in wording, but the participants indicated that the human interlocutor had shown more empathy.

“A basic statement of the interviewees was that chatbots have to become‘ more intelligent for serious contexts such as recruiting, ”they wrote FAU- Authors in their study published January 2020.

The results coincide with those of the US research institute Pew Research Center. 67% of the respondents found the idea that computer programs systematically evaluate every application without human assistance. 75% even said they wouldn't apply for a job that they knew would be hired by an algorithm.

The recruiters themselves are often faced with the so-called black box problem. The AI ​​shows you recommendations. Recommendations that are often useful and that are intended to support you in the recruiting process. However, they often cannot understand what these recommendations are actually based on. You are just there. Artificial intelligence in recruiting therefore suffers from a lack of acceptance on both sides.

On the other hand, people's basic attitudes can change quickly - this is shown, for example, by the corona crisis. Within a few weeks, many people were ready to completely change their habits - partly by necessity, but partly of their own free will. In 2019, a study by software giant Oracle and research institute Future Workplace revealed that many employees would very much welcome artificial intelligence in the workplace. 64% of respondents even said they would trust a robot more than their manager. In countries like China and India, almost 90% would even prefer a robot to their boss.

These are the biggest disadvantages of artificial intelligence in recruiting. You…

  • show no empathy
  • are not fully accepted by applicants and recruiters
  • do not always reliably eliminate biases

Does digitization make HR consultants superfluous?

"The algorithm is very good at identifying who has the necessary skills for the job on paper," said Ian Siegel in a recent television interview with Canadian financial broadcaster BNN. Siegel must know. He is the founder of ZipRecruiter, a US online job exchange that aims to bring employers and job seekers together using artificial intelligence. "He will never replace someone who speaks to another person in order to see whether they have the necessary soft skills that, together with the hard skills, decide whether they will be hired."

“It's also my job to arouse interest in a position in an initial phone call,” says personnel consultant Nikola Hänsel from Düsseldorf. Machines couldn't do that. “Personal communication is important.” The assessments are scientifically supported by the University of Erlangen-Nuremberg. “Technologies cannot completely replace communication with a person, because in some contexts (e.g. job interviews) users want a direct exchange with an empathetic counterpart,” according to their study.

The fear that algorithms will replace HR consultants in particular seems exaggerated. Especially since the personnel departments in the company have many other tasks to deal with. Compliance, taxes, data protection, personnel development and personnel retention. All of this ties up capacities. “The recruiting process takes many hours of work. This will still be outsourced in the future, ”says Nikola Hänsel. In addition, artificial intelligence does not always reduce work, but often even creates new ones. It provides more and more data and metrics that someone has to correctly evaluate, interpret and implement.

Even the greatest pessimists do not believe that HR consultants will become superfluous. Even so, a company like Randstad worries about its business model. The temporary workers want to develop from a pure recruiter to a partner, as IT boss Carsten Priebs revealed in an interview with the CIO. For example, he could imagine a "Randstad bot" that logs on to a candidate's smartphone without being asked and suggests a new job based on his current life situation. He can also easily imagine automated conversations with candidates being conducted by chatbots. The corona crisis could certainly accelerate this trend, and HR software and tools based on artificial intelligence could benefit from the pandemic.

The future belongs to the people - and the machines. In recruiting, AI works best in interaction with real people, believes Ian Siegel from ZipRecruiter. “There is nothing more powerful in the recruiting process than letting people feel that they are in demand,” said Siegel at last year's Greenhouse Open HR trade fair in New York. "Algorithms are exceptionally good at one of ten things a recruiter needs to be able to do". But nothing more. His conclusion: "AI does not take jobs away from recruiters."

Also read:

How to convince the robo-recruiter

Do you want to study machine learning? Find out more here.

Siemens wants to network the workplace environment

A contribution by:

  • Sebastian Wolking

    Sebastian Wolking is a freelance journalist and writes primarily on career topics.