Cutting corners: Large language models excel at producing grammatically correct sentences but often stumble on accuracy and clarity. Without human review, their outputs create more confusion than progress. This workslop shifts effort downstream, bogging down the very workplace processes AI is supposed to make faster and more efficient.
Modern workplaces are increasingly adopting artificial intelligence, promising speed, efficiency, and innovation. However, the reality is often messier in practice. Many companies feel pressured to adopt AI quickly, worried that failing to do so will leave them behind competitors. Yet work produced by AI can create more correction and confusion than it saves, a phenomenon Harvard Business Review (HBR) has termed "workslop."
Research from HBR's BetterUp Labs and Stanford Social Media Lab shows that AI-generated documents that appear polished can lack the substance needed to advance a task. According to Stanford's ongoing survey of US-based full-time employees, 40 percent reported receiving such outputs in the past month. Workers spend nearly two hours per incident correcting or interpreting them, creating significant hidden costs for companies. Multiplied across large organizations, those hours translate into thousands of lost workdays each year and millions of dollars in wasted effort.
Harvard Business Review cited one retail director who was less than impressed with his company's implementation of AI automation.
"I had to waste more time following up on the information and checking it with my own research," the director said. "I then had to waste even more time setting up meetings with other supervisors to address the issue. Then I continued to waste my own time having to redo the work myself."
That manager's frustration isn't an isolated case. The social and emotional toll is real. Over half of respondents said receiving low-quality AI outputs made them feel annoyed (53 percent), while nearly a quarter reported feeling offended (22 percent). Colleagues who sent such work were often seen as less capable or reliable, showing how AI missteps can ripple through team dynamics.
Even with AI adoption soaring – Gallup reports that US employees using AI at least a few times a year have nearly doubled in recent years – many pilot programs fail to generate measurable returns. An MIT Media Lab study found that fewer than one in ten AI projects delivered real revenue gains, warning that "95 percent of organizations are getting zero return" on their AI bets.
The challenge isn't just the technology itself, but how organizations deploy it. Blanket mandates to use AI everywhere often encourage mindless copy-paste behavior rather than thoughtful application. Researchers recommend clear guardrails, deliberate workflows, and leaders who set the example for using AI effectively. That can mean setting limits on where AI is appropriate, such as early drafts or routine summaries, while requiring human oversight for final outputs. When management models selective, purposeful use, employees are more likely to see AI as a tool rather than a shortcut.
Image credit: Harvard Business Review
Companies are losing money to AI "workslop" that slows everything down
