For nearly two years, Noah Olsen kept a secret from his manager at a small roofing company in Ohio. He used AI to complete about half of his software-engineering tasks, spending the rest of his time on the clock scrolling through Reddit and watching YouTube.
“I was copying and pasting all of my tasks into an AI agent such as Cursor or Claude Code, and I would let it do the work,” Olsen told Business Insider. “So instead of having to work about 40 hours a week, I would work around 20 hours.”
Olsen’s quiet use of AI to get ahead highlights a fleeting period of technological ambiguity unfolding in workplaces across the globe. With many companies just starting to adopt AI tools, workers who’ve figured out how to use them to shave hours off their workloads face an awkward dilemma: Come clean or stay mum?
Right now, there is an “arbitrage opportunity” for AI-savvy workers whose managers and peers are behind the curve, Glenn Hopper, an AI consultant in Memphis, told Business Insider.
“If you’re using AI, you’re getting polished, completed reports and spreadsheets that look incredible,” he said. “If you didn’t know AI did it, you would think someone took hours to create something like this.”
About 57% of employees said they’ve used AI at work in non-transparent ways, according to a global survey of more than 30,000 workers conducted between November 2024 and January 2025 by KPMG and the University of Melbourne. Those hidden uses included not disclosing when they used AI tools to complete their work and passing off AI-generated work as their own, the findings show.
AI’s impact on productivity can be extraordinary for some workers, said Matt Martin, cofounder and CEO of Clockwise, which uses AI to optimize workers’ calendars.
“If you’re an engineering prototyper, like, holy shit,” he said. “Your life changed in the last year.”
A recent report from McKinsey found that AI-powered agents and robots available today could technically perform about 57% of US work hours for all sorts of jobs.
Some employees are calling out the gains they’ve made using AI, said Avani Prabhakar, Chief People Officer at software maker Atlassian.
“They are gloating,” she said.
While getting work done faster is a plus, the outcome still needs to be up to snuff, Andrew Sobko, CEO of the AI infrastructure startup Argentum AI, told Business Insider. Since AI is known for making mistakes, or so-called hallucinations, users need to invest time and energy into ensuring what it spits out is accurate.
At some point in the future, Sobko said he expects enough workers to be using AI that the technology’s productivity boost will no longer go unnoticed.
“Eventually it’s going to even out,” Sobko said.
It may be a while, though, before that happens. Another recently released study from McKinsey shows that around two-thirds of companies are still in the experimentation or pilot stage of AI deployment. It also notes that companies with more than $5 billion in revenue are closer to fully embedding the technology into their operations than those with less than $100 million in revenue.
“While AI tools are now commonplace, most organizations have not yet embedded them deeply enough into their workflows and processes to realize material enterprise-level benefits,” the report concludes.
Employers hoping to boost productivity with AI should encourage workers to share the efficiencies they uncover and position that transparency as a good thing, Dan Kaplan, head of the HR practice at consulting firm ZRG, told Business Insider.
“Let’s celebrate it,” he said. “Let’s give awards for it.”
For Olsen, the software engineer from Ohio, the good times didn’t last. Over the summer, he said, his employer hired an AI specialist who taught everyone on his team to use the same shortcuts he’d been relying on. Soon after, he said he was expected to fill the hours he’d spent slacking off with additional work.
Olsen, 21 years old, quit his job in September. He then visited China for two months and is now doing freelance software-engineering work for an employer in Europe that he said hasn’t yet caught on to his penchant for AI coasting.
“If you’re not at one of the bleeding-edge companies, then you can use AI to do a lot of your work,” he said. “But who knows how long this will last.”
