HBR study: AI adoption can intensify workloads and accelerate burnout risk
New research suggests AI can expand employees’ to-do lists instead of shrinking hours, as speed and responsiveness expectations rise with adoption.
Key Takeaways
- Embedded research at a 200-person company found AI adoption can expand to-do lists and push work into breaks and evenings, even without formal pressure.
- METR reported developers using AI took 19 percent longer while believing they were 20 percent faster, signaling a planning risk from miscalibrated perception.
- An NBER study estimated about 3 percent time savings from AI adoption, with no significant changes to hours worked across occupations.
- For marketing and e-commerce teams, AI can increase output expectations unless leaders explicitly remove work and redefine responsiveness norms.
The current promise of AI at work is simple: ship faster, automate the busywork, and buy back time. New findings suggest the more teams lean in, the more work expands to fill the reclaimed minutes—creating a burnout risk that matters for marketing and e-commerce orgs optimizing for speed.
Productivity gains can translate into more work, not fewer hours
Researchers from UC Berkeley spent eight months embedded in a 200-person tech company to observe what happens when employees genuinely adopt AI tools, according to a new article in Harvard Business Review. Across more than 40 in-depth interviews, they found no formal mandate to raise targets. Instead, workers self-expanded scope: tasks that previously felt optional suddenly felt feasible, and expectations for responsiveness rose. The result was work creeping into lunch breaks and evenings as backlogs grew to match, then exceed, the time saved.
One engineer summarized the dynamic: “You just work the same amount or even more.” For B2B marketers, this can show up as “one more” landing page variant, “one more” nurture sequence, or faster turnaround on ad iterations—without reducing meeting load or review cycles.
Why AI ROI narratives can backfire for marketing and e-commerce teams
The study adds context to mixed empirical results on AI productivity. A METR experiment found experienced developers using AI took 19 percent longer on tasks while believing they were 20 percent faster, highlighting perception gaps that complicate planning and staffing (METR). Separately, an NBER paper tracking AI adoption across thousands of workplaces estimated about 3 percent time savings, with no significant change in earnings or hours worked (NBER).
For growth teams, the operational risk is that “faster content” becomes “more content,” while approvals, QA, and channel constraints remain. Leaders trying to justify AI spend may unintentionally reinforce always-on behavior.
The practical takeaway: measure AI impact as net cycle-time reduction (including review and coordination), and set explicit capacity rules—what work will be removed when throughput rises.
Stay Informed
Weekly AI marketing insights
Join 5,000+ marketers. Unsubscribe anytime.