Bosses Love AI, Workers Think It’s Useless. Who’s Right?

According to Futurism, a new survey of 5,000 white-collar workers by consulting firm Section reveals a massive chasm in how AI is perceived. A whopping 40% of non-management employees said AI saves them no time over a week, with only 2% saying it saves more than 12 hours. In stark contrast, just 2% of executives said AI saves no time, while 19% claimed it saves over 12 hours weekly. The report, highlighted by the Wall Street Journal, features workers like UX designer Steve McGarvey who says AI often gives him completely wrong solutions for accessibility problems. Furthermore, about two-thirds of regular workers feel anxious or overwhelmed by AI, while nearly 75% of executives say they are excited by it.

The Ivory Tower Problem

Here’s the thing: this isn’t just about different opinions. It’s about a fundamental disconnect in experience and incentive. Executives are hearing the gospel from people like Nvidia’s Jensen Huang, who says you’d be “insane” not to use AI for every task, or from Microsoft and Google bragging about AI-written code. They’re sold on the potential. But down in the trenches, workers are dealing with the messy reality: experimental tools that don’t fit their workflows, wrong answers that waste hours, and a constant, looming fear that this tech is meant to replace them, not assist them. When a boss says “AI will make you more productive,” an employee hears “your job is now easier to automate.” That’s a recipe for resentment, not innovation.

The Productivity Paradox

So, is AI actually boosting output? The data is… murky. That MIT study finding 95% of AI-adopting companies saw no meaningful revenue growth is a huge red flag. Other research shows AI agents failing at basic tasks and coding assistants sometimes slowing programmers down. The problem is the tech is wildly inconsistent. As UX engineer Dan Hiester noted, a task he thought would take 30 minutes consumed an afternoon with AI, while another multi-day job took 20 minutes. It completely resets how you estimate work. Basically, AI isn’t a reliable tool yet; it’s an unpredictable assistant that sometimes helps and sometimes sets your project on fire. For a manager looking at quarterly reports, the occasional big win is all they see. For the worker cleaning up the ashes, it’s a different story.

A Matter of Trust and Tools

Part of this gap might be about the nature of the work itself. An executive’s tasks—drafting emails, summarizing reports, generating broad ideas—are arguably more in the AI sweet spot. But for specialized, detailed work like writing accessible code or deep analytical tasks, current-gen AI often falls flat. And let’s be real: when leadership mandates the use of half-baked tools from on high, it destroys trust. Workers aren’t Luddites; they’ll use anything that genuinely helps. But being forced to use a glitchy, time-sinking tool because it’s the corporate fad? That’s a surefire way to kill morale and, ironically, tank the very productivity gains the bosses are chasing. The path forward requires listening to the people actually doing the work, not just the vendors selling the dream.