One woman and the real damage AI does

Y’all, I’m standing in a rabbit hole I had no intention of running down and I’m so mad I could spit. It started with an artist I admire here on Substack and damn my own curiosity because the whole mess ends where I didn’t want to go but here I am.

As if her story wasn’t plenty enough to make me angry, because it was.

First, I do not personally know Jen and she does not know me. One day there was a driftwood whale in my feed because I follow a lot of artists. Oh! How lovely. Click the heart and move on. And then another little whale and another whale, you know?

Because that’s how we mostly do, isn’t it? Life is busy.

Then she posted a note that floored me. Some butthead stole photos of her art and used AI to create “process” photos to show how “they” made them. Basically stole photos of her work and used AI to create photos to look like her work was theirs.

It happened on Facebook so she spent hours gathering proof (going back years!) and filed an an Intellectual Property claim. The end result? FB found “for” the plagiarist, issued a personal ban, and now she’s never allowed to advertise on f—book.

For. Her. Own. Work. That she’s been making for years.

And please — do not go straight to “well, Facebooks sucks” because it misses the point by a country mile and helps perpetuate the problem. I’ll explain, okay?

I am tenacious as hell when I’m angry, so I found the account that ripped her off. Want to know the long and short of it? David and Goliath.

Because she’s just one woman, and a single mom at that, walking the beach looking for driftwood, turning it into whales and dolphins, seals and bowls and selling it, get this — for charity. So when you buy a little whale, it’s helping someone who needs help and she’s helped literally thousands of people in dire need over the years.

Know how she supports her work? She runs a surprise club, where she sends out boxes and letters in the mail that arrive at your door all the way from Wales and she’s built a following of 35,000 people on Facebook who love what she’s doing.

The account that stole her work? Is a massive AI account with half a million followers. They post three DIY projects every day, which I’d guess are mostly stolen from people like Jen because every post says “AI CONTENT” at the top.

They could have chosen to ask people to participate. Hey, want to share what you’re doing? They get a post, an artist gets some views. But no. Steal. Use AI. It’s faster and when you’re in the engagement farming business, faster is good, right?

And the best part? People don’t even think about it. Just — ooh, pretty.

When I saw the photos they generated, it was dead obvious to me. The shadows and light were all wrong. Because if the light is coming from the right, the shadows fall on the left. Unless it’s AI. Then shadows fall like a drunk. Anywhere. There was one photo were the shadow was on the right for the whale, but on the left for the hand. And the props had shadows that looked like the light was from above.

Most people don’t notice stuff like that. Because “having to” notice stuff like that is as new as AI and we’re all busy, yeah? And to add insult to injury, when people saw her work on the AI account, they screamed at her that “she’s” using AI and she lost enough followers to sting. Because they accused her of using AI. Just wow.

I want to use that line from T.S. Eliot again. Not with a bang but a whimper.

Sometimes, that’s how it feels seeing this stuff.

You think it’s just art? Nah. Think again.

Here’s something you probably didn’t know. Over a year ago, Copyleaks got curious to know how much plagiarism there was in AI so they did a little test.

So you understand, I don’t mean people using AI to plagiarize. I mean AI itself.

They logged into ChatGPT and asked it to write articles on 26 topics. When they were done, they had 1,026 articles. They were 100% AI generated and they knew, because they were the ones typing in the prompts. Then they tested for plagiarism and a whopping 60% of the articles contained plagiarism. 59.7% if you don’t round up.

Direct from the belly of the beast. Plagiarism. Generated by ChatGPT.

It happened most in topics that require knowledge. Like psychology, philosophy, computer science, physics and math. Because AI isn’t human, and it can’t think and sometimes it’s easier to just grab human work to fulfill the request.

Maybe just pause a moment and think of the number of people writing about mental health, psychology and philosophy on the internet, yeah? Often, daily.

Ever wonder how some of them are churning out 7 minute reads on mental health five or seven days a week? Probably not, right? It’s like seeing where shadows fall.

Of course, that test was in late 2024, and it was ChatGPT 3.5 and we’re a few models past that now and I’m sure a lot of people will say “it doesn’t work that way anymore” but do we really believe they rebuilt the models from the ground up? They did not.

Know what’s trending today?

Basically the same thing that happened to Jen and her whales, but with words.

Here’s a thing most people don’t know. AI detection has the highest accuracy rates on pure unedited LLM-generated text. The minute you mix human work into the mix, the results are wildly inaccurate. And we’re all so busy arguing about witch hunts and Pangram errors that we don’t see the bigger picture happening around us.

Which is content theft, to be blunt.

A landmark study published by PubMed as of last year says plagiarism detection no longer works because of AI. Think about that, too. That plagiarism detection has been failing since last year.

Researchers set out to prove it, so they used AI to intentionally plagiarize several articles and then used both AI tools and plagiarism detection tools to see if the plagiarism could be detected. Short answer. No. AI makes plagiarism invisible.

According to a June 27 report at SCC Online (a legal research platform) one of the biggest problems today is that AI-driven “refurbishment” of content has normalized intellectual camouflage. Which, in human speak, means pretending something is your work when it’s not. They said any legal framework falls short to tackle the problem.

Like stolen photos of art with AI generated “progress” photos.

Or like articles and essays churned out daily from “repurposed” writing because plagiarism detectors don’t work anymore, thanks to AI rewrites.

Which means we’re all in the same boat as Jen.

Every time we write.

It makes me angry because it’s happened to me so often, I’ve lost count. And every time, the only way I knew about it was because some human reached out and sent a link and said I think this person stole your work. Is that what it’s come down to?

I don’t know what to do about it but maybe we creatives having each other’s back isn’t the worst thing. Maybe we can get Jen past 889 followers on Substack. Maybe we can keep an eye out for each other. You know? Try pay attention to what we support.

Because all I know for sure is the utter lack of respect for human work kills me a little.