Train Your Teams or Drown in Workslop

A new study published in Harvard Business Review by Stanford University and BetterUp Labs introduces a term that should worry every executive: “workslop.” That’s AI-generated work that “masquerades as good work, but lacks the substance to meaningfully advance a given task.” The study, based on a survey of 1,150 U.S. desk workers, finds that many employees are spending time decoding, correcting, or reworking their colleagues’ AI outputs rather than doing original, high-value work.

Workslop is real – and costly. The authors report that some recipients of low-quality AI content take nearly two hours per incident to clean it up. About half of survey respondents said they view coworkers who submit workslop as less creative, capable, or reliable; 42 percent said they see them as less trustworthy.

These results join a growing stack of skeptical AI ROI research. The same HBR article cites a MIT Media Lab-affiliated report showing that 95 percent of organizations see zero measurable return from generative AI pilots. That MIT “GenAI Divide” report analyzes $30–40 billion of enterprise AI spending and finds only 5 percent of custom AI initiatives reach production with measurable impact.

These reports and studies back up what we already know: AI adoption is widespread, but productivity boosts are elusive. Companies are deploying AI tools without giving teams the know-how, governance, or quality standards required to avoid downstream cleanup.

Here’s a shameless plug: if you want AI to help instead of hurt, let us help you train your workforce. Our prompt crafting workshops will help your team measurably boost productivity. If this is of interest, please visit shellypalmer.com/ai.

Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various generative AI models.