4 minBusiness
AI Productivity Gains Mask Rising Team Burnout, Managers Warn
AI tools are boosting output while hiding signs of employee burnout, according to new research and firsthand accounts from tech leaders. Managers are urged to look beyond productivity metrics and value invisible work like reviewing AI-generated code.
Artificial intelligence tools are making teams more productive, but they are also making it harder for managers to spot burnout, according to new research and firsthand accounts from technology leaders. The traditional warning sign—a drop in output—is becoming less reliable as AI helps employees maintain high performance even when they are nearing exhaustion.
A study of 442 employees at technology companies found that the link between burnout and lower productivity was weaker among those who used AI more frequently. That means managers who rely on output metrics to gauge team health may miss early signs of trouble. As one engineering leader with more than 20 years in the tech industry put it, the old indicator of slowing work no longer works when AI can keep the numbers up.
Generative AI does save time. Researchers at the University of Lausanne surveyed 83 managers and found that 82% save an average of two hours and 50 minutes per week thanks to generative AI—roughly 1.5 working days a month. But more than half of those managers spent the saved time on additional work rather than rest or development. About 36% did not even realize AI had freed up time. The result is that AI gradually raises expectations for what counts as a normal workload.
That shift is visible in how roles are changing. An architect at one tech team said he now expects junior developers to work out basic questions with an AI assistant before asking senior colleagues. A head of workforce innovation at KPMG has said she expects associates to arrive at meetings as prepared as managers once were, because AI makes that possible. These can be positive developments, speeding up growth for junior staff. But the productivity boost is not spread evenly across teams.
A case described by Sreenivasa Reddy, a lead software engineer at AT&T, illustrates the risk. His team initially became 40% faster after adopting AI. Within two months, three of their strongest mid-level engineers resigned. Junior and senior engineers were more productive, but mid-level engineers were spending more and more time reviewing AI-generated code and fixing errors. That work was not reflected in metrics or performance reviews, and they received no credit for it. The demotivating effect contributed to their departures.
This is not an isolated story. A Workday study of 3,200 managers and employees across three regions found that 85% said AI saves them between one and seven hours a week, but about 37% of that time is spent fixing and rewriting what AI has generated. That correction work often falls to middle-level roles. Because it rarely shows up in productivity statistics, the people doing it can appear less productive even as they keep the whole team’s numbers high.
Managers are being advised to look closer at what lies behind rising productivity. One recommendation is to acknowledge and value so-called invisible work—reviewing generated content and fixing mistakes. This work should count both when balancing workloads and during performance reviews. Another is to avoid turning every saved hour into more work. If AI frees up time, teams should decide in advance how to use it, whether for additional tasks, quality control, learning, or simply recharging.
Experts also warn against confusing AI use with efficiency. The real question is whether the tool helped the team make a better decision, not whether adoption numbers are rising. As one engineering leader now asks after reviewing strong results: «What did this result cost the team?» The answer may not appear on any dashboard, but it is becoming central to keeping teams healthy as AI spreads through the workplace.
6
