Every company seems determined to integrate AI, but the gains might not have a long shelf life. After an "initial productivity surge," employees who used AI reported more intense workdays and less work-life balance, and they produced lower-quality work overall, according to an ongoing study first published this week in the Harvard Business Review.

Researchers from the University of California, Berkeley, studied the habits and behaviors of about 200 people using generative AI in their work at a technology company for eight months. The company offered employees enterprise-level subscriptions to AI products. The employees weren't required to use AI, but many workers did. What happened next is exactly what AI companies hope happens: Employees who used AI worked faster and took on more responsibilities. But there were unintended consequences that showed the limits of current AI tools being used in the workplace.

One of the biggest selling points of AI in the workplace is that it can help employees handle tasks that might have otherwise been outside their expertise or skill set. Non-developers, for example, can now vibe code nearly any project. Employees in the study did this, taking on work that would have otherwise been delegated or avoided, the authors noted. So employees inadvertently created more work for themselves, putting more on their plates and struggled to balance it all.

Read More: Is AI Putting Jobs at Risk? A Recent Survey Found an Important Distinction

We also know AI as a work hack isn't without downsides. AI outputs are rarely ready to go without first being reviewed by a real human. A September 2025 study found that employees spend hours each week dealing with their colleagues' and their own low-quality or error-ridden AI work, sometimes called "workslop." A 2025 enterprise report from OpenAI said employees only saved an average of 40 to 60 minutes a week, with more time saved for power AI users.

That time saved by AI might not have made a measurable difference in work-life balance. The employees in the UC Berkeley study actually ended up working longer hours. The always-available and easy-to-use nature of AI made it simple for them to run a query during their lunch break or ask a quick question after logging off.

Even when employees had the sense of having a digital partner, their cognitive loads didn't necessarily decrease, and there were still expectations to deliver results quickly because they were using AI to help. This is why the UC Berkeley researchers said AI was more likely to "intensify" work rather than reduce it.

Authors Aruna Ranganathan and Xingqi Maggie Ye offer solutions centered around culture and norms that companies can adopt to prevent AI-powered burnout. These include protecting time for human connection, prioritizing quality results over speed, and ensuring that employees have blocked focus time without AI interruption. Being intentional with AI usage -- both in and outside of work -- is one of the best ways to prevent misuse and create work that isn't sloppy.

Across industries, workers have worried that advanced in AI will wipe out their jobs. Anthropic CEO Dario Amodei recently said AI could result in "unusually painful" short-term disruptions to the workforce. And the latest bouts of thousands of layoffs at Amazon were done explicitly because the company expected AI to fill in the gaps and help remaining employees do more with fewer resources. But we've seen ample evidence that while AI can help you do some tasks, it's unlikely to actually fulfill entire roles in most industries.