Comment on this story
Days after mass layoffs slashed 12,000 jobs at Google, hundreds of former employees flocked to an online chat room to express their sympathy for the seemingly erratic manner in which they had been suddenly fired.
They exchanged theories about how management decided who was cut. Could a “mindless algorithm carefully designed not to violate any laws” have chosen who got the axe, one person wondered in a Discord post the Washington Post has not been able to independently verify.
Google says there was “no algorithm involved” in their downsizing decisions. But ex-employees have reason to wonder, given that a fleet of artificial intelligence tools are ingrained in everyday office life. Hiring managers use machine learning software to analyze millions of employment-related data points and make recommendations on who to interview, hire, promote, or help retain employees.
Google parent Alphabet is cutting jobs, leading to more than 200,000 tech layoffs
But as Silicon Valley’s fortunes turn, this software will likely be faced with a more daunting task: helping decide who gets cut, according to HR analysts and HR experts.
A January survey of 300 HR leaders at US companies found that 98 percent of them say software and algorithms will help them make layoff decisions this year. And with companies shedding large numbers of employees — at five-figure cuts — it’s difficult for people to do it alone.
Big companies, from tech titans to companies that make housewares, often use software to find the “right person” for the “right project,” according to Joseph Fuller, a Harvard Business School professor who holds the Managing the Future of Work” co-leads the initiative.
These products build a “skills inventory,” a powerful database about employees that helps managers identify the types of work experience, certifications, and skills associated with high performers for various job titles.
The same tools can help with layoffs. “They’re just being used differently all of a sudden,” Fuller added, “because that’s where people have…a real…stock of skills.”
Staffing companies have taken advantage of the artificial intelligence boom. Companies like Eightfold AI use algorithms to analyze billions of data points pulled from online career profiles and other skill databases, helping recruiters find candidates whose applications might not otherwise show up.
Since the 2008 recession, human resources departments have been “incredibly data-driven,” said Brian Westfall, senior HR analyst at Capterra, a software review site. Resorting to algorithms can be particularly reassuring for some managers when making tricky decisions like layoffs, he added.
Lackluster earnings reports show Big Tech’s golden age is fading
Many people use software that analyzes performance data. Seventy percent of HR managers in the Capterra survey said performance was the most important factor in determining whether to fire.
Other metrics used to layoff employees may be less clear, Westfall said. For example, HR algorithms can calculate which factors make someone a “flight risk” and more likely to leave the company.
This raises numerous questions, he said. For example, if an organization has an issue with discrimination, people of color may leave the company at higher rates, but unless the algorithm is trained to know this, it might consider non-white workers a higher “risk of absconding” and suggest more of them for cuts, he added.
“You can kind of see where the snowball is rolling,” he said, “and suddenly these data points, where you don’t know how that data was created or how that data was influenced, suddenly lead to bad decisions.”
Jeff Schwartz, vice president at Gloat, an HR software company that uses AI, says his company’s software works like a recommendation engine, similar to how Amazon suggests products, helping customers figure out who to interview for open positions.
He doesn’t think Gloat’s customers use the company’s software to create lists to fire employees. However, he acknowledged that HR leaders need to be transparent about how they make such decisions, including the extent to which algorithms are used.
“This is a learning moment for us,” he said. “We have to uncover the black boxes. We need to understand which algorithms work in which way, and we need to figure out how humans and algorithms work together.”
The reliance on software has sparked debate about what role algorithms should play in job layoffs and how transparent employers should be about the reasons for job losses, labor experts said.
“The danger here is using bad data,” Westfall said, “[and] make a decision based on something an algorithm says and just blindly follow it.”
Tech workers had their free choice of jobs for years. This era is over for the time being.
But HR organizations have been “overwhelmed since the pandemic” and will continue to use software to lighten their workloads, said Zack Bombatch, an employment lawyer and member of Disrupt HR, an organization that tracks advances in human resources.
Given that, executives can’t let algorithms alone decide who to cut, and must review proposals to make sure they’re not biased against people of color, women, or the elderly — which would lead to lawsuits.
“Don’t try to blame the software on the software,” he said.