Workslop: when AI output creates more work
More and more work looks polished but adds very little once you read it. Researchers call it workslop. What it costs, why it happens and how your team can avoid it.
- reading time
- 3 min
- sources
- 3
- published
- language
- Lees dit in het Nederlands
- Workslop is AI output that looks like good work but doesn't move the task forward. The real work shifts to whoever receives it.
- In a US survey, 40 percent of office workers said they had received some in the past month.
- It costs time and trust between colleagues. Clear agreements on when and how to use AI help most.
You’ve probably seen it already. A colleague sends a two-page memo. Neat headings, bullet points, a summary at the end. Only when you read it do you realise there’s nothing in it. No decision, no numbers that hold up, no answer to the question you asked. So you follow up, dig around, rewrite. The time your colleague saved with AI has landed on your desk.
A new word for a familiar problem
In 2025, researchers at the Stanford Social Media Lab and BetterUp Labs gave this a name: workslop. They describe it as AI-generated work that passes for good work but lacks the substance to genuinely move a task forward (Niederhoffer et al., 2025).
That last part is what makes it tricky. Sloppy work is easy to spot. Workslop isn’t, because AI always produces something that looks finished.
How common it is, and what it costs
In their survey of just over a thousand US office workers, 40 percent said they had received workslop in the past month. On average, they estimated that about 15 percent of everything that lands in their inbox falls into that category. Each case cost them close to two hours of figuring out, following up or redoing.
That matches earlier research by the Upwork Research Institute (2024), in which 77 percent of employees who use AI said it had actually added to their workload, partly because they spent more time checking AI-generated content.
The hidden cost: trust
Time isn’t even the biggest issue. About half of the people who received workslop saw the sender as less creative, less capable and less reliable afterwards. More than four in ten trusted that colleague less.
That hits exactly what makes a team work. Once you doubt a colleague’s output, you start double-checking everything, and AI stops saving anyone time.
Why it happens
Workslop is rarely about bad intentions. There are usually three causes:
- People are pushed to use AI, without a clear idea of what for. So they use it everywhere, including where it doesn’t help.
- AI isn’t equally good at everything. On some tasks it adds a lot; on others people do worse with it than without (Dell’Acqua et al., 2023). If you don’t know the difference, you hand in half-finished work more often.
- Nobody feels they own the result. “The AI made it” feels different from “I made it”.
What your team can do
- Agree on what you do and don’t use AI for. Not as a ban, but as a shared choice per type of task.
- Whoever sends it, owns it. If you share something, you’ve read it yourself and can explain it, even if AI wrote the first draft.
- Ask for the point up front. What’s the conclusion, what decision is needed? Long AI texts get shorter and sharper on their own.
- Make it OK to talk about. A quick “what would you like me to do with this?” helps the sender more than quietly rewriting it.
A caveat
The workslop research is a survey: people report for themselves how often it happens and how much time it takes. That’s a good picture of how it feels, but it isn’t a precise measurement. The pattern is recognisable, though, and that alone makes it worth addressing as a team.