Recruitment 2.0: How AI Will Change How You Get Hired

Follow: , , ,

As automatization makes some jobs disappear, it’s also going to play a greater role in the hiring process. But with future jobs’ demands making the recruitment process harder, the machines may also have to become more human.

The World Economic Forum (WEF) had a nonhuman delegate for the first time this year. Attendees in Davos, Switzerland, met Hubo, a Korean humanoid robot that sparked many more or less sensationalist headlines about the end of humanity in the fight against the machine.

This worry was echoed by the latest “Future of Jobs” report recently released by the WEF, which forecasts that thousands of jobs will be lost to automatization related to the technologies of the so-called “Fourth Industrial Revolution,” including robotics, 3-D printing and artificial intelligence.

While many media reports focused on the report’s estimate that 5 million jobs will be lost to automation by 2020, less frequently mentioned were the changes artificial intelligence will make in the type of skills required for the jobs AI will create. AI and machine learning are increasingly being applied to develop tools that facilitate that hiring process, from CV screening to standing in for face-to-face interviews, affecting people’s lives to such an extent that software engineers are beginning to fret about the ethical implications of their work. While AI will make some jobs vanish, it will also help choose who is hired for those it creates.

In the past three years, companies specializing in this recruitment have emerged, largely focusing on data analysis to optimize candidate searches and job matching. Two ex-Google engineers created Connectifier, a service using big data analytics and machine learning to gather data publicly available online to create up-to-date profiles of people, which can be found even when the individual isn’t actually looking for a new job. The startup, which they call an “AI company,” has raised $6 million (U.S.) and has already grown a client base to tens of thousands of recruiters from various big companies, from Facebook to Netflix.

Another startup, which counts big tech names like SpaceX amongst its clients, is SpringRole, which matches recruiters’ requirements with candidates’ skills. “We use machine learning to match a job with a candidate profile—we look at education, employment and skills,” explains co-founder Kartik Mandaville. Recruiters can then move directly to phone interviews, having eliminated part of their manual work and saving time and money in the process.

The software is still far from perfect. Mandaville concedes that the algorithmic matching does not work equally well for all roles and functions better for those connected to the technology industry. The machine is only as good as the quantitative data that it is programmed to match, and certain jobs will have more standardized requirements.

Among the many claims for the advantages of automated recruitment, one widely held by its supporters and business observers is that, unlike recruitment by humans, it does away with possible recruiter bias. But in practice, algorithms are also vulnerable to blind spots.

“Machine-learning algorithms have the ability to pick up on biases that exist in society that we wouldn’t want to apply to the labor force,” says Nicholas Davis, head of society and innovation at the WEF and co-author of the “Future of Jobs” report. For instance, he notes that when seniority is applied as a criterion, a machine may favor traditional information showing continuous work and career progression, potentially negatively evaluating candidates who have taken parental leave or had to take time off work.

Such concerns about the implications of how a machine is programmed have led to ethical considerations becoming increasingly important in the software engineering community.

“There is still a lot of queasiness about automating the hiring process, and the fear is that we might be missing something human in it and that we might submit ourselves to a tyranny of machines,” says Julia Bossmann, founder of Synthetic Insights, a company that uses AI to accelerate progress in medical research. To address that concern, Bossmann and others are discussing the idea of a Hippocratic oath for engineers. Within the scientific community, she says, “there is a readiness to make a commitment to being on the right side of history.”

Alongside this concern among many in the field who want to ensure AI is humane is the belief that the new economy it will usher in will value more what are traditionally thought of as human skills. According to Bossmann, training in social skills will be of increasing importance, even as skills in new technologies become more widespread, particularly coding.

“If everyone only learns to set up a goal and follow steps ABC to get there, that may lead to a lot of inter-human conflict,” Bossmann says. “The human skills and the people skills and the openness to consider other people’s perspectives and interact with them are important complementary subjects to learn,” she says, adding that these skills are becoming more important even when recruiting for engineering positions.

“Complex problem solving, the ability to have critical thinking [and] positive interpersonal relationships in the work setting have become more important relative to other skills,” says Davis. His WEF study found that most large-scale employers expected the emphasis to shift toward these “emotional” skills in the future economy. These sorts of people skills are also potentially more enduring than computing skills, which are quickly outdated as technology races ahead.

“The other skills are empathy and engagement—putting yourself in the shoes of another person and understand not just relationships but how to turn those relationships into value for communities and organizations,” says Davis. It’s an irony that even as machines play a greater role in hiring, the next revolution in HR may focus on emotional intelligence, which is often relatively neglected in today’s recruitment processes.

For the time being, most AI technology in recruitment remains focused on data analysis, facilitating preinterview stages of selection. But some are already predicting AI will move into the traditional person-to-person steps of the hiring process—video interviews analyzed by a computer for facial expressions or keywords, for example.

“A guided automated interview would probably be a more effective way of assessing skills than it would be using what we have traditionally now,” says Davis.

Such ideas can make one wonder if the future of recruitment might come to resemble the Voight-Kampff machine test from Philip K. Dick’s novel “Do Androids Dream of Electric Sheep?” (the inspiration for Ridley Scott’s film “Blade Runner”). In the book, the system was used to help police determine whether an individual was human or android.

This time around, it’s likely to be the machine asking the questions.

Leave a Reply

Your email address will not be published. Required fields are marked *