It finally happened to you: you asked your coworker a question and they gave you an answer that’s obviously copy-pasted AI output. Or maybe they’ve sent you a document they’ve “written”, but it reads like AI slop. Or (for software engineers), maybe your coworker just started leaving comments on your code that look like unfiltered AI output.
You’re starting to get annoyed, but you aren’t sure what to do. Should you tell them off and cause a confrontation? Send them a link to noslopgrenade.com or dontpastetheai.com and come off as passive-aggressive? Politely try to hint that maybe you were hoping for a different answer?
I’ve run into this situation a few times and I think I have an approach that works reasonably well.
Warning: this may not work for everyone
Admittedly, this approach works best if you’re the person’s manager. And you should probably avoid it if the copy-paster is a thin-skinned person who holds a lot of power over you.
Between those two extremes, you’re more likely to succeed if you’re in a more senior position, or have accrued credibility with the person, or if your coworker has signaled that they’re very open to feedback.
First, identify the actual problem
Ask yourself: what specifically about the incident caused you inconvenience?
Was it that you had to read through a wall of text when a single sentence would have been enough, wasting your time and attention? Was the document confusing to read and follow? Did you have to wade through 10 comments on your code that obviously missed the context? Is it that you were hoping for their own opinion but now you aren’t even sure they’ve read the question?
Sometimes copy/pasted AI output isn’t actually a problem. For example, a couple of weeks ago, a co-worker of mine sent me a comment on my code that said “Claude Code says: {quotes Claude Code}. Do you think it’s worth checking out?” On the surface, that should have been annoying, but this coworker (1) is often right, and (2) is usually thorough in his work and reviews. So this comment was likely filtered from many others left by AI and probably contained a valid concern.
But suppose there’s an actual problem, and you’ve identified what it is.
Next, bring this up with the coworker
I usually open the conversation with something like:
Hey {name}, are you open to some unsolicited feedback on {thing in question}?”
Or some variation of that. This does a couple of things:
The answer will probably be “yes”, and hopefully as a result they’d be at least partly open to the feedback.
You’re setting up an expectation that there’s some negative feedback coming so it doesn’t just drop on them unexpectedly.
Next, follow the rules of constructive feedback:
Be direct; don’t dance around the problem.
Concentrate on the behavior and its negative effect on you rather than concentrating on the other person.
If possible, prefer to use “I” statements like “I was confused…”, “I had to spend more time…” rather than accusatory “you” statements like “you made me waste time”. The feedback might be negative enough as it is, and phrasing it as “you’re bad” is an unnecessary extra slap in the face. Of course, sometimes it’s unavoidable.
Keep in mind that it’s not the use of AI in general that’s the problem, it’s this specific use that’s the problem.
So your next message could be something like:
When you give me unfiltered AI output as PR comments, I have to spend time figuring out whether each of the comments even makes sense. It would be a lot more helpful if you pre-filtered the AI findings yourself and left only those that you think actually apply.
Or,
I found that the document was confusingly organized: {add specific problems}, and I had a hard time following it. I think it would have been a lot clearer if it had… {specific suggestions}. In general, I find that AI is pretty bad at organizing documents, and I have to edit them myself afterwards.
The consequences
I’ve had various responses from coworkers to this. One was very open to feedback, said thank you, and moved on. One was more defensive, but the feedback was eventually internalized and the issue did not reoccur.
Whatever happens, don’t get into an argument—it will only entrench the other person further in their views. You’ve left the message; you can move on now.