I found a post by Gordon McLean called "When Is AI Image Slop Ok?" and I want to answer it straight: no. Slop's not ok. That doesn't mean AI images are off the table, I use them on this blog myself sometimes, tied to whatever I'm actually writing about, chosen on purpose. What's not ok is the lazy, thoughtless kind that gets passed off as real. Not in a blog post, not in a newsletter, not anywhere with your name on it.
Gordon wasn't defending slop either, to be fair. He was working through his own use of AI for featured images on his site, going back and forth on whether that counts as the same kind of slop everyone complains about, since it's "just" a banner image nobody studies closely. But one comment did more work than the whole rest of the post. A reader named Rhys Wynne pointed out that if you can't tell whether the image is real or AI-made, why would anyone trust the words underneath it? Gordon came back the next day with a follow-up saying he was done using AI images, full stop, going back to real photos from sites like Unsplash instead. That's the whole argument, really. It was never about the image. It's about trust.
I already use Unsplash myself for photos on this blog. Reading Gordon's follow-up made me realize I should start adding a photo credit caption under them too, giving credit to the actual photographer instead of just dropping the image in with no source.
I've circled this idea a few times already on this blog. AI as a tool to help you think, organize a mess of notes, or get a draft moving is fine by me. AI as a stand-in for the actual choosing and care that goes into making something is not. Slop is what happens when nobody decided anything, no image was picked, no thought went into what it should look like, it just got generated because that was faster than choosing something real. A slop image on a post looks small by itself, but it's a tell. It says the work wasn't worth real effort, so why would a reader owe it real attention back?
A featured image is the first thing a reader sees. It sets the tone before a single word gets read. If that image was generated in five seconds with no thought behind it, it's already telling the reader something about what the rest of the post is worth. And once someone starts wondering if the picture is fake, they start wondering about everything else too.
There's a difference worth naming between making a new image with AI and using AI to alter something real. What crosses a real line is taking an actual photo and twisting it into something that didn't happen, then passing it off as real. Trump has done exactly that more than once. The White House posted an altered version of a real photo of Homeland Security Secretary Kristi Noem that added her crying, which isn't what the original photo showed. Trump also shared a doctored photo of Barack and Michelle Obama walking off an Air Force One covered in graffiti, a scene that never happened, and the graffiti included messages that got the image called out as racist right away. Those aren't jokes. They're real photos edited to plant a fake memory in people's heads.
I get that mocking politicians with doctored images isn't new. People have done it for centuries, with cartoons and photo tricks alike. But that history doesn't make it fine now, and it's not only a political thing either. The same move shows up all over the internet.
Fake animal photos dressed up to look like real wildlife shots. Fake movie posters made to look like a real announcement for a movie that doesn't exist. I see that one constantly scrolling Facebook, and I end up stopping to double-check almost every time. That gets tiring fast, and it shouldn't be something I have to do in the first place. Pictures that use kids to make some made-up story feel more real so people share it before checking. All of it works the same way: borrow the trust people give to something real to sell something that isn't real at all. None of it is ok.
There's one more piece to this that goes past just images. When AI shows up in news or on social media, the speed is what makes it dangerous. Something gets posted, it looks close enough to real, and people believe it before anyone stops to check. Experts who study this call it an erosion of trust, and that's the right word for it. Once a fake image or clip spreads far enough, correcting it later barely matters. The fake version already did its damage in people's heads. That's what worries me more than any single picture: not that AI can fake something, but that so many people will believe it at sight and never think to check.
So no. AI image slop is not ok. Not a little ok, not "depends on the context" ok. Not ok.