AI and other technologies must be balanced with human interaction
By Brian Kreissl
One of the themes I touched on last week was the ever-increasing desire to leverage technology in the HR profession. Since we’re hearing so much about artificial intelligence (AI) these days, there is definitely some pressure to use these types of technologies in managing the workforces of the future.
While HR doesn’t want to appear to be Luddites or look like we’re not up to the task of using current technologies to streamline processes and enhance efficiency, responsiveness and insight relating to the workforce, I believe there’s a darker side to some of these technologies. We must also be careful not to take the “human” out of human resources.
Problems with artificial intelligence
One potential problem with AI was recently highlighted when online retailer Amazon found their new recruitment tool was biased against women. Apparently, the software started preferring men for software development and other technical roles and had literally “taught” itself to prefer male candidates over female ones.
One of the exciting and interesting concepts relating to artificial intelligence is machine learning, where computers actually start to “learn” things based on previous experience. However, in the case of the Amazon recruitment tool, the software learned to have a bias in favour of men simply because a larger proportion of males had previously applied for jobs in the technology field.
Apparently, the tool penalized applicants who had the word “women’s” in their resumés. It also developed a bias against candidates who had attended two all-women’s colleges. While some reports have said Amazon recruiters didn’t base their decisions solely on the tool and the company itself said the tool was never used to evaluate candidates, this still illustrates at least two potential problems with artificial intelligence.
One problem is obviously, without proper parameters in place, machine learning can result in computers learning undesirable things. Another issue is the problem of “garbage in, garbage out,” with imperfect algorithms resulting in suboptimal output.
I mention this story not to have a go at Amazon, which I’m sure never set out to discriminate against female candidates, but to show some of the potential perils of using AI for recruitment.
While HR professionals are being told that the future will belong to AI, that human decision-making will be enhanced in the workplace through such technologies and if we don’t learn to use them we will be left behind as a profession, there’s also the very real fear of AI leading to massive job losses.
While technology up until now has created many opportunities and created just as many jobs as it destroyed — many of which were more highly-skilled and better paid than the ones they replaced — AI appears to be something different. When computers and robots become self-programming and self-correcting, there may be a time in the future when human beings are largely redundant to the economy.
I worry about things like self-driving cars and trucks leading to job losses among truck drivers and other people who drive for a living. What will all those people do if they’re thrown out of work? I also believe we’d be foolish to think AI couldn’t result in significant job losses even in white collar high-touch professions like HR that until now have relied on human interaction to add value.
Replacing HR professionals with technology
If we try to replace human beings with technology in the human resources function, I also believe we’re likely to see a backlash from employees and jobseekers and a negative impact on organizations’ employer and talent brands.
While the technology is improving, how often, for example, do we encounter problems with automated attendants and interactive voice response (IVR) systems when calling a contact centre? Often it seems almost impossible to get the system to understand what you want or route your call to a live person. That kind of frustration can be encountered, for example, when contacting an HR call centre about one’s benefits package.
There is also widespread frustration with applicant tracking systems (ATS) that reject people who are perfectly qualified for the roles they’re applying for. Many people are also starting to rail against the one-way video recordings some organizations are starting to include as part of the application process.
While such technologies can create efficiencies and streamline processes, as a profession we must do our best to add value through “high touch” interaction with human beings.