Workslop
Workslop is AI-generated or AI-assisted workplace material that appears polished enough to hand off but lacks the context, accuracy, judgment or substance needed to advance the task. Its defining practical effect is transferred effort: a recipient must interpret, verify, correct or redo the output. The label is narrower than AI slop and does not cover every imperfect draft made with AI.
Origin and context
Six authors introduced the label in Harvard Business Review on 22 September 2025. The associated BetterUp Labs and Stanford Social Media Lab page says the findings came from an online survey of 1,150 full-time U.S. desk workers conducted that September; 40% reported receiving workslop in the prior month. This supports joint research attribution rather than Stanford alone. It was a survey-based HBR publication, not a paper titled 'AI Slop at Work' as the base record states.
Why it matters
The concept shifts attention from a sender's local speed to team throughput. A generated memo can be quick to produce yet force a colleague to reconstruct assumptions and evidence. Microsoft Research adopted the same recipient-burden framing in its 2025 technical report on the future of work. Zapier later used 'AI workslop' in a separate survey of 1,100 U.S. enterprise AI users. Because Zapier measured time spent revising AI outputs rather than the original receive-and-handoff construct, it documents circulation and a related burden, not a replication of the original prevalence estimate.
Example
An analyst sends a polished forecast whose figures lack sources and whose assumptions do not match the project. A colleague must recover the inputs and rebuild the analysis; that handoff fits workslop. An explicitly labeled rough draft that the team expects to refine is not necessarily workslop. Productivity theater instead means behavior intended to signal busyness or availability, such as unnecessary meetings or visible status activity. It can be entirely non-AI and need not transfer an artifact, although an AI-generated document can serve both patterns.
How it differs
AI slop
AI slop is the broader category of low-quality generative content across public and private settings. Workslop is its workplace-specific form, centered on a seemingly usable handoff and the cognitive or corrective burden shifted to another worker.
Automation bias in agentic AI
Automation bias is a person's overreliance on automated recommendations despite better contrary information. Workslop describes an output and handoff problem. A recipient may accept workslop because of automation bias, but either condition can occur without the other.
Maturity and evidence
Skills Intelligence rates workslop at maturity 3 with an established lifecycle. Its origin is traceable, Microsoft Research uses the same core meaning, and Zapier independently applied the label in empirical workplace research. The term remains below maturity 4 because it is only about a year old, has no standardized measure, and current studies do not operationalize the boundary identically. Longitudinal, cross-country research using a shared definition would justify a higher rating.
Limits and open questions
Neither reviewed survey demonstrates that workslop causes organization-wide productivity loss. BetterUp and Stanford sampled U.S. desk workers online and relied on respondents' recognition of the label; Zapier sampled U.S. AI users at larger companies, used unweighted results and included revision of one's own output. Their percentages are therefore not directly comparable. Workslop is also an evaluative label: assess the artifact against the task, the recipient's context and the actual rework required rather than inferring low quality from AI use alone.
Related terms
References
- AI-Generated ‘Workslop’ Is Destroying ProductivityHarvard Business Review · 2025-09-22 · class B
- Workslop: The Hidden Cost of AI-Generated BusyworkBetterUp Labs · 2025 · class A
- Microsoft New Future of Work Report 2025Microsoft Research · 2025 · class A
- Most workers spend 3+ hours per week cleaning up AI workslopZapier · 2026-01-14 · class B
- WTF is productivity theater?WorkLife · 2023-04-26 · class B
- Exploring automation bias in human–AI collaboration: a review and implications for explainable AIAI & Society · 2025-07-03 · class A
Last updated: 2026-09-07