Generative AI Boosts Productivity But Leads to Burnout, Study Finds

Got story updates? Submit your updates here. ›

A study by UC Berkeley researchers Aruna Ranganathan and Xingqi Maggie Ye found that employees who were offered free access to enterprise generative AI tools ended up working faster, taking on more tasks, and working longer hours - often without being asked. While this initial boost in productivity was observed, the researchers also documented negative effects like burnout, cognitive fatigue, and weaker decision-making over time.

Why it matters

The findings challenge the common narrative that AI will reduce employee workloads, highlighting the potential for AI to actually increase demands on workers if not implemented thoughtfully. As more companies mandate AI adoption, this study provides important insights on how to mitigate the risks and protect employee wellbeing.

The details

Ranganathan and Ye conducted an 8-month study observing a 200-employee tech company that provided free access to generative AI tools. They found that workers took on a broader scope of tasks, worked at a faster pace, and worked longer hours, even without being asked to do so. The researchers attributed this to AI lowering the barriers to entry for many tasks, allowing employees to 'squeeze small amounts of work' into breaks and transitions. This led to more multitasking and a more continuous involvement with work, rather than natural pauses. Over time, the initial productivity boost gave way to burnout, cognitive fatigue, and weaker decision-making.

• The study was conducted from April to December 2025.

What they’re saying

What’s next

The researchers recommend that employers implement more intentional pauses or dedicated break periods to help employees regulate their work tempo, as well as sequencing norms to shape when work moves forward with AI assistance. They also advise companies to avoid letting employees become too isolated from human interaction and collaboration.