Becoming an AI Editor: Asking, Is this any Good?
OpenAI released GPT-5 last week. A lot has been written about it already and whether it lives up to the hype of being compared to a planet-destroying super-weapon that can be blown up with a single shot to its reactor core. So I’m not going to write about that.
Instead, I’m going to follow up on editing AI-created content. I’ve previously addressed how you need to read content to be able to take responsibility for the output and edit it to deal with large language models’ mythomania (which GPT-5 has not solved). Today, we’re going to address a bigger question: evaluating if what you have in front of you is any good.
For the sake of this piece, I’m going to define “good” as “meeting its purpose.” Your step here as an AI editor is making sure that what’s been created is actually what you needed in the first place.
So, first, take a moment to remember. Why did you write your prompt in the first place? Were you trying to create a piece for education? To persuade an audience to a point of view? To entertain yourself or your friends? To sell something?
From that perspective, read critically, and think of some basics of good writing from the perspective of your audience.
If educating: Make sure you’ve meeting your audience where they’re at. Does what’s written assume knowledge your readers might not have? Fill those gaps.
If persuading: Read to ensure the argument actually has a logical succession of ideas. Is anything introduced out of context or out of a sensible flow.
If selling: Make sure you’re addressing your audience’s pain points, your solution’s benefits that address those pain points, and that you have a clear call to action.
To entertain: Is it at least amusing to you? Are there any obvious errors that would take someone out of the experience you’re trying to create for them?
To get something off your chest: Do you agree with what’s being said? If someone said this to you across a cup of coffee, would you be concerned they were having a stroke?
From there, watch out for LLM-isms. If you aren’t reading analyses of LLMs’ writing constantly, it may take a little time to catch on to these. My least favorite is pointless juxtapositions: “It’s not x, it’s y” or “this isn’t about x, it’s about y.” Hyperbole is also endemic to the species.
From there, you’ve done your best. Like writing by yourself, writing with an LLM and editing it is a skill, and it’s one that changes with each model upgrade. But if you’re reading and editing carefully and conscientiously, you can avoid the embarrassment of not being able to answer for content that you’ve overseen. And as an editor, that’s one of the bars we’re trying to clear every day.
Happy editing.