Across offices and shop floors, artificial intelligence is being sold as a pressure valve for overloaded workers. Yet the lived reality for many people looks more like a treadmill that keeps speeding up, as every efficiency gain is swallowed by rising expectations, new tasks and constant oversight of machine output. The data now suggests that AI is not simply failing to ease work, it is actively reshaping what “enough” looks like in ways that leave employees more exhausted.
Instead of freeing people from drudgery, AI is often treated as a reason to cram more into the same eight hours, or to quietly stretch those hours into nights and weekends. That shift is colliding with already alarming levels of burnout and stress, turning AI from a promised safety net into another weight on workers’ backs.
The burnout baseline: a workforce already at its limit
Any honest look at AI’s impact has to start with the fact that workers were already running hot before chatbots and copilots arrived on their desktops. Research summarized under the banner Job Burnout At shows burnout at 66% in 2025, with employees reporting that they simply have more work than time to do it and that this strain is feeding labor shortages and turnover. A companion analysis from New Study Shows reinforces that picture, noting that a sizable share of staff are stressed specifically because their task lists outstrip the hours available.
Against that backdrop, it is striking how quickly leaders have pinned their hopes on automation as a cure. A recent video discussion framed around the question AI Driving Productivity highlights that 96% of senior executives expect AI to boost performance, even as they acknowledge rising stress. When I put those numbers side by side, the pattern is clear: leaders are betting that technology will fix a workload problem that is, at its core, organizational and cultural. That mismatch sets the stage for disappointment on both sides.
AI exhaustion and the rise of “workslop”
Inside many companies, the first wave of AI deployment has not been surgical or strategic, it has been chaotic. One analysis of the AI Paradox of reports that 83% of companies say new AI systems are actually worsening workplace stress instead of relieving it. A related breakdown from Jul describes “AI exhaustion” as a growing risk to what it calls a company’s most valuable asset, its people, when tools are bolted on without redesigning processes or expectations.
On the ground, that exhaustion often takes the form of what one study labels an “interpersonal workslop tax.” When a colleague uses a generative system to crank out a messy draft, others must spend time deciphering, editing and sometimes redoing that workslop before it is usable. I see this dynamic most clearly in knowledge roles, where AI-generated slide decks, reports or code snippets look complete at a glance but hide subtle errors that only surface after hours of review. The net effect is that the apparent productivity of the person who hit “generate” is subsidized by the invisible labor of everyone who has to clean up after the machine.
The productivity paradox: more tools, more stress
There is now a growing body of research showing that piling on AI tools does not automatically lighten the load. A survey of 2,500 workers, summarized under the line Adds To The, found that many employees feel AI has increased both their workload and their anxiety, with some considering quitting within months because they feel overworked. Another analysis of The Productivity Paradox notes that employees with more AI tools actually report higher stress, challenging the assumption that more automation equals less pressure.
Academic work is picking up the same signal. A study summarized under the heading While AI finds that while AI can speed up certain tasks, it also introduces new cognitive demands, from monitoring outputs to managing complex interfaces. A related piece on Researchers: AI’s Productivity underscores that the promised time savings often come bundled with new forms of mental load. Put simply, the tools may be fast, but the expectations they unleash are faster.
When “good” keeps moving: expectations in the AI era
One of the most useful metaphors I have seen for AI’s impact on work comes from a comparison to household appliances. A widely shared reflection notes that when washing machines arrived, they did not give people back their evenings, they changed what counted as acceptable cleanliness so that Instead of washing occasionally, families expected Clothes to be spotless all the time. The same pattern is playing out at work: once AI can draft ten versions of a proposal in minutes, managers start to expect ten options, not one carefully considered version.
That moving target helps explain why burnout is rising even as tools proliferate. Commentary grouped under the line AI isn’t reducing argues that AI mostly redefines what “good” looks like, rather than shrinking the pile of work. I see that in how performance reviews now quietly factor in whether employees are “leveraging AI,” and in how job postings increasingly treat AI proficiency as a baseline requirement rather than a bonus skill.
Frontline workers: relief with strings attached
The story is more nuanced on the frontline, where AI is often embedded in scanners, scheduling apps or inventory systems rather than in text generators. Research summarized as Frontline Workers Who finds that frontline staff who use AI heavily are less likely to report burnout and are often more optimistic that AI will help them. A separate report on the Impact AI on these workers notes that those using AI are less likely to feel overwhelmed, even as concerns about job security linger.
Another analysis of how AI use can help Frontline staff beat burnout points out that this segment represents roughly 80% of the global workforce, so even modest improvements in their experience matter enormously. Yet the same research cautions that frontline burnout remains high, which suggests AI is offering partial relief rather than a cure. I read this as evidence that when AI is tightly integrated into specific tasks, like automating stock counts or optimizing routes, it can genuinely remove friction, but it does not erase low pay, unpredictable schedules or safety risks that drive much of the stress.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
