Study Reveals AI Tools Intensify Workloads Rather Than Reducing Them

A recent UC Berkeley study suggests that AI tools may be increasing workers’ burdens instead of lightening them, creating a phenomenon researchers call “workload creep.”

AI’s Impact on Workplace Productivity

Researchers from UC Berkeley’s Haas School of Business conducted an eight-month study of a tech company with 200 employees to examine how AI affects workplace dynamics. Their findings, published in the Harvard Business Review, reveal that AI tools often lead to increased workloads rather than the promised efficiency gains.

One employee summed up the experience: “You had thought that maybe, oh, because you could be more productive with AI, then you save some time, you can work less. But then really, you don’t work less. You just work the same amount or even more.”

Key Findings of the Study

The research identified several concerning patterns in how AI adoption affects workers:

• Employees voluntarily adopted AI tools, initially excited by the potential for increased productivity
• Workers began absorbing tasks they would normally outsource or require additional help for
• Engineers spent more time correcting AI-generated code from colleagues
• Increased multitasking as employees ran AI tools in the background while working manually
• Work began to infiltrate free time as employees used AI during breaks and meetings
• Blurred boundaries between work and personal time reduced the rejuvenating quality of downtime

The Vicious Cycle

The researchers identified a troubling pattern: AI accelerated certain tasks, which raised speed expectations. Higher speed made workers more reliant on AI, which expanded the scope of what they attempted, further increasing both the quantity and density of work.

This cycle ultimately led to fatigue, burnout, and lower quality work—the opposite of what AI proponents promise.

Broader Context

The Berkeley Haas study adds to mounting evidence challenging the AI industry’s productivity claims:

• An MIT study found most companies adopting AI saw no meaningful revenue growth
• Research shows AI agents frequently fail at common office tasks
• Other studies document how AI can produce low-quality “workslop” that colleagues must fix
• 40% of non-management white-collar workers report AI saves them no time

Looking Forward

While the researchers suggest companies should establish stronger guidelines for AI use, the study reveals how difficult it is to manage the negative effects of AI implementation. The technology can easily produce unintended consequences that are challenging to address and still being discovered.