Unilever's use of new screening software raises a few questions
It was just over a year ago when I wrote the story “Is AI biased in recruitment?” for Canadian HR Reporter. The article looked at Amazon’s apparent misstep when it abandoned a recruitment tool that turned out to be biased against women.
The computer program had issues around gender, along with too many unqualified candidates. It screened applicants by looking at patterns in resumés submitted to Amazon over 10 years, but most of these came from men so, essentially, the system taught itself that male candidates were preferable, according to Reuters.
For the December issue, I’m working on another story about AI and recruitment. This time, it’s about consumer goods giant Unilever using data points including facial expression, body language and keywords during job interviews in the United Kingdom.
Through this AI, the company was apparently able to filter up to 80 per cent of the candidate pool, according to HireVue Assessments, provider of the video interviewing software and pre-hire assessments.
The new process has meant not only improved recruits, but big savings for Unilever along with reduced recruiting time, it said.
But soon after the announcement was made, critics were questioning the approach.
Without an “extremely diverse dataset,” the system will be biased, according to Griff Ferris, legal and policy officer for Big Brother Watch, in the Telegraph.
“Using a faceless artificial intelligence system to conduct tens of thousands of interviews has really hilling implications for jobseekers,” he said. “This algorithm will be attempting to recognize and measure the extreme complexities of human speech, body language and expression, which will inevitably have detrimental effect on unconventional applicants.”
Of course, there’s been plenty of coverage of the downsides to facial recognition technology, particularly when used for security purposes. Some U.S. cities have gone so far as to ban government use of the technology because of concerns about its accuracy and privacy.
If a monster company like Amazon had issues with the technology, it’ll be interesting to see why Unilever has such confidence in this particular robotic approach.
One big reason for the appeal of AI in recruitment could be the demands of recruitment, and finding recruiters. LinkedIn recently released a report saying the ever-growing scarcity of top talent has increased demand for recruiting professionals by 63 per cent globally.
That number jumps to 166 per cent in Canada, with just over one quarter (26 per cent) of Canadian recruiting professionals saying that retaining top talent recruiters is a priority of over the next five years.
Workers just aren’t staying in their jobs as long as they used to. They’re more comfortable trying out a job for a couple of years and then moving on, instead of moving up the corporate ladder with one or two employers.
With that higher turnover, recruiters are that much busier trying to keep up.
And with more people from different demographics looking for a bigger variety of perks on the job — whether that means stock options, working from home or more vacation days — recruiters and employers are scrambling to figure out what exactly attracts and retains the top candidates.
Does that mean AI is the answer? While I’m OK with the concept of software scanning through resumés for keywords that match jobs — as long as the algorithm is covering every possible scenario and not missing out on potential future stars — there’s something about technology scanning people’s faces as they react to interview questions on video that just doesn’t sit right.
If I was an employer, I’d be nervous about what exactly is being programmed into the system as desirable from candidates and, more importantly, the technology’s’ capability at actually getting that right.