LEE COUNTY, Fla. — Artificial intelligence has been around for years, but the technology is being used in new ways, including to monitor your performance on the job.
It may sound like something out of a movie, but that’s exactly what happened to employees at Amazon. Florida Gulf Coast University professor Chrisann Ruehle said the software’s criteria was simple.
“They track how long people are working on their screen, they track keystrokes. They’re tracking a lot of performance-related data,” Ruehle said. “In addition, they’re also tracking the amount of time that people are spending on breaks, the amount of time people are spending idle.”
If the artificial intelligence didn’t like the data, it would flag the employee.
“The system had the ability to flag those individuals. And several years ago, the company did not have a lot of human oversight,” she said. “Rather, once the bot identified this gap, people were not meeting performance expectations, if they did it three times or more, the bot actually terminated the employee.”
That led to a class-action lawsuit in 2018. Ruehle said there are many reasons why a worker might not hit the metrics. She also said the metrics may not be the best way to tell if work isn’t getting done.
“Perhaps there are health issues that are going on, perhaps there are internet access issues,” she said. “There are still many people around the country that have issues with accessing the internet. They may be able to accomplish their work, the same quality of work, in a much shorter time frame. So in that case, if they are still providing a high-quality product, end product, why should we penalize them for that?”
With more companies using artificial intelligence for all kinds of tasks, Ruehle said it’s important that managers have some say over who stays and who goes.
RELATED STORY: Experts say artificial intelligence is coming into the workplace faster than previously thought
“The human oversight is being able to take that information and have a one-on-one dialogue,” she said. “We cannot use technology as a substitute for managing those relationships with workers.”
Instead, Ruehle said managers have to talk to their workers and take AI data into consideration.
“That information would be sent to the supervisor,” she said. “The supervisor then should take that information and they need to have a conversation with the employee, and they need to figure out.”