Research: The Concealed Drawback of Employing AI at Work

0:28This episode is brought to you by Google Chrome. of america for a reason with suvs made to move with your rhythm the versatile equinox tackles your entire day the spacious traverse fits your crew and your whole weekend and tracks bring style with value you can count on all infused with tech that has your back so your drive always hits the right court Chevrolet together let's drive two years after CHPT launched only 16 % of American workers are using AI at work even though 91 % are allowed to most people assume that gap is about skill or training but our research suggests something deeper people are afraid that using AI will make them seem less competent we call this the competence penalty and it's quite a derailing AI adoption inside companies.

1:20I'm Ozacar and in my recent HBR article, based on research led by my collaborators Phyllis Giagayi, Yang Ping Tu, and Giayi Hu, we explored this phenomenon across a global tech company. In the study, over 1 ,000 software engineers reviewed the same piece of code, but in some cases they were told it had been written with the help of AI, and in others by a person alone. The results were striking. Engineers rated the person who used AI as less competent, even though the code never changed. That's the competence penalty. And it doesn't impact everyone equally. Women face much harsher judgment than men.

1:57And the harshest critics? Engineers who hadn't adopted AI themselves, especially men. This fear of being judged leads many employees to quietly avoid AI, even when it could make their jobs easier. This was clear in follow-up surveys with nearly 1 ,000 engineers. Those who feared the competence penalty the most were also the least likely to adopt AI. That fear was especially common among women and older workers, groups already navigating bias in tech environments. The result? Companies lose out on productivity and reinforce inequality. At one tech company in our study, low adoption alone led to an estimated 2.5 % to 14 % loss in annual profit.

2:39To break the competence penalty, try these three steps. 1. Map the hotspots. Identify teams where AI adoption is low and social risk is high, like when women or older employees are outnumbered by non-adopting male peers. 2. Spotlight visible role models. Encourage respected leaders, especially those from underrepresented groups, to use AI openly. 3. Rethink Evaluations. Remove AI-used tags from code reviews. Focus on outcomes, not methods, and use blind reviews to reduce bias. If you want to lead a successful AI transformation, start by making your culture safe for everyone to participate. You can read the full article at hbr.org.