What do you say when a collaborator sends you AI slop to edit?
You can’t accuse them of AI slop, because you never really know.
But the thing is, we know that readers are suspicious of slop and will stop reading if they suspect that, whether its warranted or not.
So here’s a live list of coded ways I have of telling people to reduce the AI slop in their writing. Or maybe its human slop, either way it needs to go back to the author for more editorial work.
On lists: What is the argument you are trying to make here? Can you rephrase to make that point.
On contrastive language (‘not x, but y’): Tell us directly what you did do, not what you didn’t do.
On adjectives and adverbs: Avoid adjectives and adverbs in scientific writing.
On restatement: This conclusion restates the intro. Describe how your paper has advanced research in this topic.
On overuse of ‘critical’ and similar adjectives: Say how its important rather than relying on adjective to do the work.
On meaningless content: Is this sentence specific to your study, or could it be used in a study on another topic? If so, go back to the topic sentence of this paragraph and refine what you want to say.
On 5000 word drafts of discussions (I’m seeing this a lot!): This is long, let’s go back to the outline. Send me the topic sentences for each paragraph so we can work on relevance and flow, then you can move content over that’s relevant and cut unnecessary content.
Over-use of em-dashes… Now these are controversial. You could say to switch out for some colons instead, but I think the issue is often deeper. Probably the final point above ‘go back to the topic sentence of this paragraph and refine what you want to say’ will be helpful as well. I’ll think about it next time I see one in the wild.
And if your paper mentions goblins then I’m not reading it at all (unless its about goblin sharks!).