Executives promise AI productivity gains while employees drown in "workslop"

Companies increasingly sell AI as an instant efficiency booster, but for many office workers it does the opposite: instead of saving time, they get a stream of neatly formatted but half-baked material that they have to rewrite by hand. The effect already has a name — "workslop": AI-generated work that looks convincing at first glance but contains errors, inaccuracies and lost context. The core problem isn't that people are lazy. It's that businesses demand more content and faster decisions than their processes can adapt to.

One such story comes from a copywriter at a large Miami-based cybersecurity company who spoke under the pseudonym Ken. After layoffs, management required the remaining staff to use chatbots, citing productivity gains. Drafts did start appearing faster, but the team then had to spend even more time rewriting them, correcting factual errors and sorting out contradictions between answers from different AI tools. According to him, quality dropped noticeably, turnaround times grew and team morale suffered. When employees pushed back against the imposed strategy, management blamed them for the decline in efficiency.

Surveys confirm the gap between what executives expect and what workers experience. In one study of 5,000 US office workers, 40% of rank-and-file employees said AI saves them no time at all, while 92% of senior executives believe the opposite. Another study, conducted by BetterUp with the Stanford Social Media Lab among 1,150 knowledge workers, found that 40% had encountered "workslop" at least once in the past month. Cleaning it up takes several hours a month on average, and for a 10,000-person company the productivity losses can run into millions of dollars a year. In other words, AI looks like an accelerator in slide decks and like a new layer of hidden bureaucracy in day-to-day work.

The problem isn't limited to marketing or copywriting. Designers complain that colleagues paste unedited bot responses into chats and emails and then can't explain what they actually meant. In medical clinics where staff were pushed to generate patient replies with AI, the time savings also proved questionable: clinicians report extra editing, frustration and data security risks.

Researchers attribute this less to weaknesses in the models themselves than to poorly defined tasks. Many companies roll out AI as a do-everything tool without clear rules, training or lines of accountability. At the same time, some employers are already trying to justify their generative AI investments by cutting headcount, even though the returns remain far from what was promised. By frequently cited estimates, most companies still see no meaningful return on their AI investments.

The takeaway is uncomfortable but important: deploying AI does not by itself make a company more efficient. If a company first lays people off, then tells those who remain to "speed up with bots," and never changes its quality control processes, it gets not automation but a cascade of rework and eroding trust within the team. For businesses, this is a signal to judge AI not by slick demos and upbeat executive reports, but by how much time goes into fixing the output, where errors occur and who is accountable for the final text, design or customer reply.

And until companies have clear AI usage rules and genuinely involve employees in those decisions, "workslop" risks becoming not a side effect but the new normal of office work.