Harvard study reveals AI's hidden cost: Employee burnout and unsustainable workloads

A Harvard Business Review study has uncovered a troubling paradox in AI adoption: while generative AI tools make work feel more accessible and rewarding, they're also driving employees to work longer hours with greater intensity, often at the cost of their wellbeing.

Researchers Aruna Ranganathan and Xingqi Maggie Ye from UC Berkeley's Haas School of Business studied 40 workers at a 200-employee tech company between April and December last year. Their findings paint a picture of AI as a "fun new toy" that quickly transforms into an "unrelenting taskmaster."

The productivity trap

The study revealed that employees using AI tools worked at a faster pace, took on broader scopes of tasks, and extended their workdays into evenings and early mornings. Critically, this wasn't mandated by the company—workers did it voluntarily because AI made "doing more" feel possible and intrinsically rewarding.

However, this productivity surge came with significant costs. The intensity of work, longer hours, and expanded responsibilities left employees feeling stretched too thin as added work encroached on personal time. Researchers warned that "workload creep can in turn lead to cognitive fatigue, burnout, and weakened decision-making."

The hidden costs of AI acceleration

The study identified several concerning patterns:

• Task expansion: Employees began taking on responsibilities that previously belonged to other roles, boosted by AI's cognitive assistance
• Quality control burden: Engineering staff found themselves checking novice coders' work, coaching "vibe coders," and finishing others' projects
• Blurred boundaries: The ease of starting tasks led workers to engage during breaks, nights, and early mornings with fewer natural pauses
• Cognitive strain: What appears as productivity gains in the short term masks growing mental fatigue as employees juggle multiple AI-enabled workflows

"Because the extra effort is voluntary and often framed as enjoyable experimentation, it is easy for leaders to overlook how much additional load workers are carrying," the researchers noted.

The broader context

This study emerges against a backdrop of conflicting predictions about AI's impact on employment. While Forrester estimates AI could eliminate 6% of jobs (about 10.4 million) by 2030 through various automation technologies, the question of whether this translates to genuine productivity gains remains contested.

J.P. Gownder, Forrester's vice president and principal analyst, recently expressed skepticism about AI revolutionizing productivity, suggesting the promised gains may not materialize as expected.

Solutions for sustainable AI adoption

The Berkeley researchers propose several strategies to prevent AI-induced burnout while maintaining productivity:

Intentional pauses

Organizations should implement "decision pauses" that require, before finalizing major decisions, one counterargument and one explicit link to organizational goals. This "widening the attention field" helps prevent drift from core objectives.

Structured sequencing

Rather than demanding continuous responsiveness enabled by AI, teams should work in coherent phases, moving deliberately rather than at AI's maximum pace. "Make the team lead the AI; don't let the AI lead the team," they advise.

Human connection preservation

As AI enables more solo, self-contained work, organizations must protect time for human interaction. Brief check-ins, shared reflection moments, and structured dialogue can interrupt continuous solo engagement with AI tools and help restore perspective.

The warning signs

The researchers caution that the fear of job loss may drive employees to overwork themselves with AI tools. With predictions of significant job displacement looming, workers may feel compelled to demonstrate maximum productivity, even at the expense of their health.

"For workers, the cumulative effect is fatigue, burnout, and a growing sense that work is harder to step away from, especially as organizational expectations for speed and responsiveness rise," the study concludes.

This research suggests that the AI revolution's true challenge may not be technological implementation but rather managing the human costs of accelerated productivity. Organizations that fail to address these issues risk not only employee wellbeing but also the quality of work and long-term sustainability of their operations.

The study serves as a crucial reminder that technology adoption must be accompanied by thoughtful human-centered policies to prevent the very tools designed to enhance productivity from becoming instruments of worker exhaustion and organizational dysfunction.