Why does AI-made work create more work?
When AI saves you an hour and costs your colleague two.
A colleague sends you a doc. Six pages, neat headings, a confident summary at the top.
Twenty minutes in, you realise it doesn't actually say anything you can act on. There are no real numbers and no decision, just a lot of well-formatted words.
Now you have to work out what they meant, or do it again yourself.
Workslop: AI-made work that looks finished but doesn't move the task forward.
Researchers at Stanford's Social Media Lab and BetterUp Labs coined the term in 2025. They described it as material that "may look impressive, but contains very little in the way of actionable facts and figures."
In their survey of US workers:
It flows in every direction. Staff send it to managers, and managers send it to staff.
The work didn't disappear. It moved to the person who received it.
Merriam-Webster picked slop as its 2025 Word of the Year:
digital content of low quality that is produced usually in quantity by means of artificial intelligence.
Simon Willison has a simpler test: slop is content that's unrequested and unreviewed. Nobody asked for it, and nobody checked it before sending it on.
Polish used to take effort, so a polished document usually meant someone had thought hard about it. Now polish is free (see "The nicer AI's work looks, the less we check it").
So the sender skims it, sees that it looks right, and hits send. The checking moves from the person who made it to the person who reads it.
Read every line. If you haven't read it, don't send it.
Cut it to what the reader needs. Shorter is kinder. Two clear paragraphs beat six pages.
Say what you did. "Drafted with Claude. I checked the numbers against the March report." That one line tells the reader how far to trust it.
Don't quietly redo it. Ask for what's missing:
Thanks. What's the one decision you need from me, and where do the numbers come from?
That's quicker for you, and it teaches the sender what useful looks like.