Artificial intelligence boosters often promise the tech will lead to a reduced workload. AI would draft documents, synthesise information, and debug code so employees can focus on higher-value tasks. But according to recent findings, that promise is misleading.
An ongoing study, published in the Harvard Business Review, joins growing bodies of evidence that AI isn’t reducing workloads at all. Instead, it appears to be intensifying them.
Researchers spent eight months examining how generative AI reshaped work habits at a US-based technology company with roughly 200 employees. They found that after adopting AI tools, workers moved faster, took on a wider range of tasks, and extended their work into more hours of the day, even if no one asked them to do so.
Importantly, the company never required employees to use AI. It simply offered subscriptions to commercially available tools and left adoption up to individuals. Still, many workers embraced the technology enthusiastically because AI made “doing more” feel easier and more rewarding, the researchers said.
That enthusiasm, however, came with unintended consequences. Over time, workloads quietly expanded to overwhelming levels. The gradual, often unnoticed, creep in responsibilities led to cognitive fatigue, burnout, and weaker decision-making.
While AI can produce an initial productivity surge, the researchers warn that it may ultimately contribute to lower-quality work and unsustainable pressure.
To track these changes, the researchers observed the company in person two days a week, monitored internal communication channels, and conducted more than 40 in-depth interviews across engineering, product, design, research, and operations. They found that job boundaries began to blur.
Employees increasingly took on tasks that previously belonged to other teams, using AI to fill knowledge gaps. Product managers and designers started writing code. Researchers started handling engineering tasks. In many cases, work that might once have justified additional hires was simply absorbed by existing staff with the help of AI.
For engineers, the shift created a different kind of burden. Rather than saving time, they spent more hours reviewing, correcting, and guiding AI-generated work produced by colleagues. What had once been straightforward code review expanded into ongoing coaching and cleanup of flawed outputs.
The researchers described a feedback loop: AI sped up certain tasks, which raised expectations for speed. Higher expectations encouraged greater reliance on AI, and that, in turn, widened both the scope and volume of work employees attempted. The result was more activity, not less.
Many participants said that while they felt more productive, they did not feel any less busy. Some actually felt busier than before AI arrived. “You had thought that maybe, oh, because you could be more productive with AI, then you save some time, you can work less,” one engineer told the Harvard Business Review. “But then, really, you don’t work less. You just work the same amount or even more.”
What looks like a productivity breakthrough, the researchers concluded, can actually mask silent workload creep. And overwork, they warn, can erode judgment, increase errors, and make it harder for organisations to distinguish genuine efficiency gains from unsustainable intensity.
To counter these risks, the researchers proposed a protective approach they call “AI practice,” a set of intentional norms and routines that define how AI should be used at work and, crucially, when to stop. Without clear boundaries, they caution, AI makes it easier to do more but harder to slow down. – Inc./Tribune News Service
