Workers Don’t Want Bosses Knowing They Use AI—Even As They ‘BYOAI’ To Work

CEOs are talking a lot about generative AI, calling it a transformative technology and hoping it will turbocharge productivity. Just one problem: Many workers, worried they’ll look replaceable, are afraid to admit they use it.

Just over half of employees who use AI at work—52%—say they’re reluctant to divulge they’re applying it to their most important tasks, and roughly the same number worry that using it on critical tasks in their job makes them look replaceable, according to the latest Work Trend Index from Microsoft and LinkedIn. Still, workers are clearly using the new tech: Some 75% of full-time office workers surveyed said they’re now using AI at work—up from 46% six months ago—and more than three-quarters are turning to their own tools rather than company-provided ones, a phenomenon Microsoft calls “BYOAI.”

The annual report, released Wednesday, draws from a survey of 31,000 full-time workers, as well as LinkedIn labor and hiring data and usage data Microsoft captures from customers using its software. While the survey was commissioned by Microsoft—which sells its Copilot AI tool and has a big stake in artificial intelligence firm OpenAI—it still offers a snapshot of how workers and companies are using generative AI and thinking about it as the tech shifts from shiny new object to a tool that’s quickly reshaping how work gets done.

“The first thing we really see in this report is that employees want AI at work and they’re not waiting for companies to catch up,” says Colette Stallbaumer, who leads the company’s future of work and Microsoft 365 teams. “A lot of that’s unsanctioned use.”

This year marks the first time Microsoft has worked on the report with LinkedIn, which it acquired in 2016, and found a growing role for AI in both job titles and in what jobseekers and managers expect. The position of “head of AI,” the analysis found, has tripled over the past five years and grew by more than 28% last year. Meanwhile, LinkedIn has seen 142 times more members add AI skills to their profiles, the report says, while job posts that mention AI have seen applications grow 17% over the past two years compared to listings that don’t mention those skills.

By “AI skills,” neither LinkedIn—nor the people who add them to their profiles—are necessarily referring to highly technical machine learning skills. “We’re so early in this that it’s all just getting called AI skills,” says Aneesh Raman, a vice president and workforce expert at LinkedIn. “Understand what it is, and then really start playing around with it. You’ll realize it is not some highly technical skill set.”

The report also shares findings from Microsoft’s software users. Researchers designed a randomized, six-month trial of 60 customers across industries that it calls the “first mass-scale observation of 3,000 individuals” using AI at work without interventions. The big takeaway: Copilot users spend less time on email—they read 11% fewer messages and spent 4% less time on them, Microsoft says—and edited more documents in applications like Word or Excel. Interestingly, the trial found that some Copilot users spent more time in meetings—perhaps because they have more time to meet with colleagues due to efficiency—while others spent less time in them.

But the hesitancy the survey found about employees revealing they use AI puts numbers to a phenomenon that has concerned other executives and researchers. At some companies, employees may fear sharing they use ChatGPT or another AI chatbot because there’s a workplace policy against its use. But at many others, it’s more about a mindset shift away from employees being able to take credit for their work at a time when many are concerned about job security amid a shifting economy and a perceived threat of AI.

Ethan Mollick, a professor at the University of Pennsylvania’s Wharton School who studies AI, calls this “shadow IT.” Employees, he told Forbes in a recent interview, “know if they tell you about it, then maybe you’re not so impressed by all the amazing work they’re doing,” he says. “Everyone is trying to not show that they’ve automated their work.”

In his recent book, Co-Intelligence, Mollick writes that research shows people judge work differently when they know content has been produced by AI rather than a person. As a result, “much of the value of AI use comes from people not knowing you are using it,” he writes in his book, mentioning an unscientific social media poll he conducted where AI users said they don’t reveal they’re using it at least some of the time. “If someone has figured out how to automate 90 percent of a particular job, and they tell their boss, will the company fire 90 percent of their co-workers? Better not to speak up.”

As a result, businesses that want to reap efficiency from AI may want to spotlight workers who use it as a way to reassure them it’s valued, and add training to highlight that it’s an expected part of the job. Microsoft’s report finds that just 39% of people using AI say they’ve gotten training from their company. “Companies need to have a strategy [for AI]. They need to have a point of view,” says LinkedIn’s Raman. “This is change management at a whole new level.”