AI promises productivity. Faster outputs. Higher value work. More time for strategic thinking.
But what if the opposite is quietly happening?
A study from the University of California Berkeley, featured in Harvard Business Review, tracked what actually happens when employees are given access to AI tools. The findings challenge the assumption that AI automatically reduces workload.
In many cases, it intensified it.
What the Researchers Found
Researchers Aruna Ranganathan and Xingqi Maggie Ye followed 40 employees at a 200-person US tech company over eight months. The staff worked across engineering, product, design, research and operations.
AI tools were introduced. They were not mandatory. No one was forced to use them. Employees chose to experiment.
What happened next is telling.
Instead of using time savings to focus deeply on one important task or to switch off earlier, employees worked faster and took on more. They expanded their responsibilities. They extended their hours into early mornings, evenings and even lunch breaks.
Nobody asked them to.
AI made doing more feel possible. Even satisfying. So people quietly did more.
The Disappearing Pause
One of the most striking patterns was how AI filled the gaps in the workday.
Employees began sending prompts to AI right before meetings or lunch, letting the system draft documents or analyse data while they stepped away. On the surface, that looks efficient.
But over time, those micro tasks eliminated natural pauses. People checked AI outputs during breaks. Late at night. First thing in the morning. Because it was so easy to start “just one more” AI-assisted task, the boundary between work and rest began to dissolve.
Light experimentation slowly became expanded responsibility.
“Workload Creep”
The researchers call this pattern workload creep.
Once AI tools were available, employees started stepping into tasks that would normally belong to someone else or that might have justified hiring another person. They corrected colleagues’ work, fed meeting notes into AI for summaries, drafted additional reports and absorbed adjacent responsibilities.
AI made those cross-boundary tasks feel manageable.
Over time, jobs expanded. Roles did not. Pay did not. Headcount did not.
Many participants reported feeling stretched too thin as additional tasks began eating into personal time.
Why This Matters Beyond One Study
In the short term, it looks like a productivity surge. More output. More projects. Faster turnaround.
But the Berkeley researchers warn of silent strain. Workload creep leads to cognitive fatigue, weaker decision-making and burnout. Eventually, quality drops and the early gains disappear.
A separate study from Harvard Business School reinforces an important point. When asked which jobs should be automated, people supported automating about 30 percent of roles AI could currently handle. Even when told to imagine a future system that outperforms humans at lower cost, resistance remained, especially in roles where human presence matters.
Efficiency is not everything. People value trust, connection and judgement.
Organisations that optimise purely for output risk damaging both employee well-being and customer relationships.
The Guardrails That Matter
The Berkeley researchers are not anti-AI. They are pro-deliberate management.
Their recommendations are practical.
Build intentional pauses.
Before finalising major decisions, require teams to articulate at least one counterargument and clearly link decisions to organisational goals. Reflection prevents blind acceptance of AI-generated output.
Sequence projects deliberately.
Just because AI can accelerate everything at once does not mean it should. Move forward at a sustainable pace.
Protect human connection.
Short check-ins, shared reflection, and structured dialogue interrupt the pattern of constant solo engagement with AI. Without these pauses, people operate in continuous output mode without questioning quality.
Their core principle is simple. The team should lead the AI, not the other way around.
My Take
This study should be required reading for every manager rolling out AI tools.
The risk is not that AI fails.
The risk is that it succeeds too well. That productivity increases mask rising fatigue quietly. Those roles expand without redesign. That people burn out before leadership notices the cost.
AI is a brilliant tool. But a tool without guardrails becomes a treadmill.
If your organisation is deploying AI, ask yourself this.
Are you measuring what people produce?
Or are you also measuring what it costs them to produce it?
Those are two very different questions.
I’m Joanne Sweeney. If you want to build your AI literacy and skills, I can help.
📚 AI Playbook is available now.
📰 Subscribe to Human in the Loop, my LinkedIn Newsletter
🎓 The CPD Accredited AI Playbook Diploma is open for enrolment.
📞 And if you want to talk about AI training for your organisation, book a call at digitaltraining.ie
Sources
UC Berkeley Haas / Harvard Business Review: “AI Doesn’t Reduce Work, It Intensifies It”
https://hbr.org/2026/02/ai-doesnt-reduce-work-it-intensifies-it
The Register: Coverage of the Berkeley study
https://www.theregister.com/2026/02/11/ai_makes_employees_work_harder
Entrepreneur: “AI Is Making Employees Work More, Not Less”
https://www.entrepreneur.com/business-news/ai-is-making-employees-work-more-not-less/502600
Harvard Business School: “People Are Mostly OK With AI Taking Over Many Jobs, Up to a Point”
https://www.library.hbs.edu/working-knowledge/riley-performance-resistance-2026
About the Podcast Host
AI and Communications Specialist | Author | Podcaster | International Speaker
Joanne Sweeney is a recognised leader in digital marketing and digital transformation, specialising in the practical application of AI to enhance communication strategies. The founder of Digital Training Institute, Joanne, empowers leaders and communications professionals to leverage AI responsibly and effectively, ensuring transparency, trust, and tangible results.
She completed the Harvard Kennedy School Executive Programme in AI and Public Policy this summer before launching her fourth book AI Playbook, in September. Joanne has also published two editions of Public Sector Marketing Pro, which were preceded by her first title, Social Media Under Investigation Law Enforcement and the Social Web.
Her books have become go-to resources for governments, public sector and policing agencies, as well as marketing and communications professionals, all navigating the challenges of the changing Internet.
Joanne is the host of the AI SIX Podcast, a daily six-minute podcast that provides actionable insights, expert interviews, and practical advice to help professionals navigate the new world of work.
