In a national survey released last month by Stanford University, AI-generated “workslop” is having a growing negative impact on workplace productivity. This phenomenon, defined in a Harvard Business Review article published on September 27, 2025, refers to outputs from artificial intelligence that are often “unhelpful, incomplete, or missing crucial context.” The new, continuous survey has collected responses from 1,150 full-time U.S. workers. It illustrates that nearly 40% of these respondents experienced workslop in the last month.
Workslop points to an increasing fear — even among AI’s biggest proponents — that what the machines create, while perhaps technically proficient, won’t be effective or trustworthy. As organizations begin to adopt AI tools as part of their business processes, concerns around the quality of output have taken center stage. Our survey findings showed that workslop is a major quality killer. It builds unnecessary hurdles for staff as they work to aggressively address tasks that are unfinished, poorly defined, or both.
In almost every instance, workslop results in additional work burden on grantees. As stated by researchers from BetterUp Labs, “The insidious effect of workslop is that it shifts the burden of the work downstream, requiring the receiver to interpret, correct, or redo the work.” This back-and-forth can create a toxic cycle of rework that further reduces productivity and stifles team efficiency.
Our ongoing survey examines the scale and effects of workslop in more detail. Stakeholders looking to get involved can find it on Stanford University’s Qualtrics link. Their findings would be instrumental in helping us understand how organizations can more effectively allow for and control AI-generated outputs.
In light of these challenges, BetterUp Labs researchers recommend that organizations “model thoughtful AI use that has purpose and intention.” They suggest that leaders “set clear guardrails for your teams around norms and acceptable use.” Taking these actions will make it possible to bring AI tools much further into the fold. In doing so, they’ll minimize the harmful impact that results from abuse of those tools.
These issues with workslop point to a wider discussion on AI’s place in the workplace. As organizations increasingly embrace these technologies at scale, reckoning with workslop will be critical to upholding the standards of productivity we hope to achieve.
