A growing number of employees are quietly using AI to reduce their workloads without informing their managers, taking advantage of a window of technological uncertainty in the workplace. Noah Olsen, a 21-year-old software engineer from Ohio, used AI tools like Cursor and Claude Code to complete about half of his tasks at a small roofing company, allowing him to work only 20 hours a week instead of the standard 40. He spent the rest of his time browsing Reddit and watching YouTube, all while maintaining the appearance of productivity. Olsen’s experience reflects a broader trend: as AI tools become more powerful and accessible, employees who master them gain a significant edge. Glenn Hopper, an AI consultant in Memphis, described this as an “arbitrage opportunity” — workers using AI to produce high-quality results in far less time than traditional methods would allow. If the output looks polished and professional, managers may assume it took hours to create, not minutes. A global survey of over 30,000 workers conducted by KPMG and the University of Melbourne between November 2024 and January 2025 found that 57% of employees had used AI at work in ways that weren’t transparent. This included hiding the use of AI to complete tasks or presenting AI-generated work as their own. The productivity gains can be dramatic. Matt Martin, CEO of Clockwise, which uses AI to manage work schedules, said the impact on engineering and prototyping roles has been transformative. A McKinsey report estimates that today’s AI agents and automation tools could technically handle 57% of all U.S. work hours across various job types. Still, not all AI output is flawless. Andrew Sobko, CEO of AI infrastructure company Argentum AI, warned that AI often produces errors or “hallucinations,” requiring careful review and correction. Workers must still invest time to verify accuracy — meaning the time savings aren’t always as dramatic as they seem. Eventually, Sobko predicts, widespread AI adoption will level the playing field. Once most employees are using AI, the productivity advantage will disappear. But that shift hasn’t happened yet. McKinsey’s latest findings show that around two-thirds of companies are still in early stages of AI experimentation or pilot programs. Larger firms with over $5 billion in revenue are ahead of smaller ones, which often lack the resources to integrate AI deeply into workflows. Experts like Dan Kaplan of ZRG suggest employers should encourage transparency. Rather than punish employees for using AI efficiently, companies should celebrate it — rewarding innovation and sharing best practices. “Let’s celebrate it,” Kaplan said. “Let’s give awards for it.” For Olsen, the window of opportunity closed when his company hired an AI specialist who trained the entire team on the same tools he’d been using. Suddenly, he was expected to fill the extra hours with more work. He quit in September, traveled to China for two months, and now works freelance for a European employer that hasn’t yet noticed his reliance on AI. “I’m not sure how long this will last,” he said. “But if you’re not at a cutting-edge company, you can definitely use AI to do a lot of your work — for now.”