Imagine this: you’ve just discovered a tool that can complete in five minutes what used to take you an hour. Instead of feeling liberated, you find yourself working longer hours, managing multiple workflows simultaneously, and ending each day mentally exhausted. This isn’t a dystopian prediction – it’s happening right now in offices across America, and it’s reshaping how we think about artificial intelligence’s impact on work.
The Unintended Consequences of Agentic AI
New research from the University of California Berkeley reveals a surprising trend: rather than reducing workloads, AI tools are making work more intense. In a study of a US tech company, researchers found workers extending their hours into early mornings and evenings while taking on broader responsibilities. Most strikingly, these changes happened organically – the company didn’t mandate AI use or request longer hours.
The Berkeley researchers identified three dynamics driving this intensification. First, AI’s ability to fill knowledge gaps encourages workers to take on tasks outside their expertise. Second, the ease of starting new tasks means workers fill every gap between meetings with additional work. Third, the ability to hand off tasks to AI agents leads to constant multitasking, with workers managing multiple workflows simultaneously.
“The knowledge that another five minutes of back-and-forth with agentic AI can set in train a task that would have taken me an hour or more has meant I suddenly want to set in train another task, and another,” writes John in the Financial Times’ AI Shift newsletter, describing his personal experience with what LSE professor Luis Garicano calls “a version of the Jevons Paradox for work effort.”
The Corporate Pushback and Resistance
While some workers are voluntarily intensifying their work, corporations are taking a more direct approach to ensure AI adoption. Accenture, the global consulting giant, has implemented a controversial policy linking promotions to leadership positions with regular AI tool usage. The company monitors senior employees’ weekly log-ins to tools like AI Refinery and SynOps, facing what executives describe as an exercise in “chivvying” older senior figures who are often less comfortable with technology.
Accenture CEO Julie Sweet previously stated the firm would “exit” staff who couldn’t adapt to the AI age, and the company has trained over 550,000 people in generative AI. However, the policy has faced criticism, with some employees calling the tools “broken slop generators” and others threatening to quit if affected. The approach highlights the tension between corporate AI mandates and employee adoption challenges.
The Broader Economic Impact
Beyond individual workplaces, AI’s expansion is creating macroeconomic ripple effects. A Goldman Sachs report reveals that AI’s soaring electricity demand is fueling inflation, crimping consumer spending, and slowing economic growth. Electricity prices rose 6.9% last year – more than twice the Federal Reserve’s preferred inflation measure – and data centers’ share of US electricity consumption has roughly doubled since ChatGPT’s 2022 rollout.
Goldman Sachs estimates that higher electricity prices will lower consumer spending growth by 0.2 percentage points on average in 2026-2027 and exert a 0.1 percentage point drag on GDP growth, with lower-income households most affected. This creates a paradox: while AI promises productivity gains, its infrastructure demands may undermine economic growth.
When AI Creates More Problems Than It Solves
The challenges extend beyond work intensity and economic impacts. In UK workplaces, HR teams are drowning in lengthy, AI-generated employee grievances that often include irrelevant historical details, incorrect legal precedents, and made-up legislation. New employment tribunal cases increased by 33% in three months to September compared to a year earlier, while concluded cases decreased by 10%.
“I suspect that AI is behind it,” says Anna Bond, legal director at Lewis Silkin. “The length of complaints about working conditions, colleagues and managers is the most pernicious problem.” This trend highlights how poorly implemented AI can create administrative burdens rather than reduce them.
Finding Balance in the AI Revolution
The Berkeley researchers recommend practical steps to secure AI’s benefits without incurring its costs. If compulsive AI use is eating up unplanned breaks and moments of reflection, these may need to be reintroduced in structured form. Managers may need to shift from checking that work is progressing well to ensuring it’s not going too fast.
Canadian computer science professor Margaret-Anne Storey suggests that human sign-off on AI-generated changes should involve not just what was changed, but how and why – ensuring teams retain full understanding of projects. Meanwhile, platforms like RentAHuman represent an intriguing counter-trend: an online marketplace where AI agents hire humans for tasks requiring physical presence or human skills, suggesting new employment opportunities rather than just displacement.
As Sarah notes in the Financial Times discussion, “Professional office-based work lends itself to this sort of behavior in a way which isn’t necessarily true in other work settings.” She calls this the “suitcase principle of white-collar work” – similar to how you always fill your suitcase whether going away for a week or fortnight. The question becomes: are we filling our workdays with meaningful output or just more busy-work?
The transition to AI-enhanced work requires more than just tool adoption – it demands a fundamental rethinking of work rhythms, management practices, and success metrics. As companies and workers navigate this new landscape, the challenge isn’t just implementing AI, but doing so in ways that enhance rather than exhaust human potential.