Sarah thought she’d hit the productivity jackpot when her company introduced AI writing assistants last spring. The marketing manager could now draft client proposals in half the time, turning what used to be three-hour marathons into quick 90-minute sprints. She felt like a superhero with a digital sidekick.
But six months later, Sarah found herself checking emails at 9 PM and revising presentations on Sunday mornings. Instead of leaving the office earlier, she was staying later to tackle the growing pile of “quick tasks” that AI made seem effortless. What started as relief had somehow become a burden.
Sarah’s experience isn’t unique. A groundbreaking study reveals that AI tools are creating more work for employees, not less, fundamentally changing how we think about artificial intelligence in the workplace.
The Paradox of AI Productivity
Researchers at Berkeley Haas School of Business spent eight months tracking 200 employees at a major tech company, and their findings challenge everything we thought we knew about AI in the workplace. While employees using AI tools completed individual tasks faster, they ended up working longer hours and taking on significantly more responsibilities.
- French Army Receives 1,000 Sonora Drones in Record Time—Here’s What That Means for Defense
- Poland quietly considers A-29 Super Tucano jets to counter growing drone threats
- Canada could become home to massive Gripen fighter jet assembly line under Saab’s surprising new proposal
- Poland’s quiet A330 MRTT aircraft purchase could change everything for NATO’s eastern defense strategy
- France’s Fury ALM drone combines missile speed with AI precision in ways that could reshape warfare
- Air Force quietly restocks its most devastating weapon after Iran strikes underground nuclear sites
The study, conducted from April to December 2025, gave workers access to ChatGPT-style AI assistants without requiring their use. What happened next surprised even the researchers studying workplace behavior.
“Employees worked faster, took on more tasks and stretched their working day — often without being asked to and without realizing the long-term cost,” explained Dr. Michael Chen, the study’s lead researcher.
The phenomenon researchers call “work intensification” showed up in three distinct ways. Workers completed existing tasks more quickly, expanded into new types of work they previously wouldn’t have attempted, and increasingly worked during personal time because AI made tasks feel manageable anywhere.
Why AI Extra Work Feels Like Personal Choice
Perhaps the most troubling discovery was how voluntary this extra workload felt to employees. Nobody received memos demanding longer hours or additional projects. Instead, workers genuinely believed they were choosing to do more because AI made it easier.
The study revealed several psychological factors driving this behavior:
- AI tools create a false sense of task simplicity, making complex projects seem manageable
- Faster completion times generate dopamine hits that encourage taking on more work
- Fear of being left behind pushes employees to demonstrate AI proficiency through increased output
- Blurred boundaries between AI assistance and personal capability lead to overconfidence
- Managers inadvertently reward quantity over quality, reinforcing the cycle
“The most dangerous part is that workers feel empowered, not exploited,” noted workplace psychologist Dr. Amanda Rivera. “They’re making what feels like autonomous decisions to work more, which makes it harder to recognize the problem.”
The research team documented how AI tools gradually shifted workplace expectations. Tasks that once justified full days now seemed like quick afternoon projects, creating pressure to fill the “saved” time with additional work.
| Work Aspect | Before AI Tools | After AI Implementation |
|---|---|---|
| Average Daily Tasks | 6-8 tasks | 12-15 tasks |
| After-Hours Work | 2 hours/week | 7 hours/week |
| Project Complexity Range | Limited to expertise | Expanded beyond specialty |
| Weekend Work Sessions | Rare | Common |
The Hidden Health Cost of Always-On Productivity
While productivity metrics soared, employee wellbeing quietly deteriorated. The Berkeley study found that workers using AI tools reported higher stress levels, despite feeling more capable and efficient during individual tasks.
The research revealed a concerning disconnect between short-term satisfaction and long-term sustainability. Employees felt accomplished completing more tasks, but struggled with persistent fatigue, difficulty disconnecting from work, and a sense that they could never do “enough.”
Physical symptoms emerged after just three months of intensive AI use. Workers reported eye strain from extended screen time, neck pain from poor posture during extended work sessions, and sleep disruption from late-night “quick tasks” that stretched longer than expected.
“We’re seeing a new form of digital burnout,” warned occupational health specialist Dr. James Martinez. “Workers feel productive and in control, which masks the traditional warning signs of overwork.”
The study documented increased sick days among heavy AI users, despite their higher productivity ratings. Managers initially dismissed this as coincidental, but patterns became undeniable after six months of data collection.
What This Means for Workers and Companies
The implications extend far beyond individual productivity. Companies celebrating AI-driven output increases may be unknowingly creating unsustainable work environments that could lead to mass burnout and talent retention problems.
For individual workers, the research suggests several immediate concerns. The boundary between work and personal time becomes increasingly blurred when AI tools make tasks feel effortless. Many employees find themselves responding to work requests during family dinners or vacation days, convinced they’re making quick, harmless choices.
Labor experts worry about the long-term impact on workplace standards. If AI-assisted productivity becomes the baseline expectation, workers without access to advanced tools may face impossible performance standards.
“We’re potentially creating a two-tier workforce where AI access determines who can meet basic job expectations,” observed labor economist Dr. Patricia Wong.
Companies must also grapple with the legal and ethical implications of AI-driven work intensification. While employees voluntarily work more hours, employers benefit from unpaid overtime disguised as enhanced efficiency.
The study recommends several protective measures for both workers and organizations:
- Establish clear boundaries for AI tool usage during off-hours
- Monitor workload increases alongside productivity metrics
- Train managers to recognize signs of AI-enabled overwork
- Create policies preventing task expansion without corresponding compensation
- Regular wellbeing checks for employees using AI tools extensively
Moving forward, the challenge isn’t limiting AI tools but learning to use them sustainably. The Berkeley study suggests that without conscious intervention, artificial intelligence may create more problems than it solves, turning workplace relief into an exhausting illusion of enhanced capability.
FAQs
Why do AI tools lead to extra work instead of saving time?
AI makes tasks feel easier and faster, encouraging workers to take on more projects and extend their working hours voluntarily, creating a cycle of increased workload.
Are companies forcing employees to work more with AI?
No, the study found that workers voluntarily choose to work more because AI tools make additional tasks seem manageable and rewarding.
What health problems are linked to AI-intensive work?
Researchers documented increased stress, eye strain, neck pain, sleep disruption, and higher sick day usage among heavy AI tool users.
How can workers protect themselves from AI-driven overwork?
Set clear boundaries for AI use during personal time, monitor your actual working hours, and resist the urge to take on extra tasks just because AI makes them seem easy.
Should companies limit AI tool access to prevent overwork?
Rather than limiting access, companies should establish policies around workload increases and monitor employee wellbeing alongside productivity metrics.
Will this problem get worse as AI tools improve?
Potentially yes, as more sophisticated AI tools may make even complex tasks feel simple, further expanding the scope of voluntary overwork.