Have you ever run your own writing through an AI detector? It's an eye opener.
Have you ever run your own writing through an AI detector? It's an eye opener.
You'd be surprised how many people write like a robot.
Happy Friday,
Have you ever put your own writing into an AI detector? Until this week, I hadn’t. I mean, why would I? I know I wrote it.
Earlier this week I ran across a thread where freelance writers were talking about clients running their writing through AI detection and refusing to pay them.
Some of them swore up and down they did not use any AI.
Wow. Ouch. That has to hurt.
If they didn’t use AI and got stiffed, I feel for them. That sucks.
Good grief, I can’t imagine doing the sweat and research to write a 750 word blog post or 1500 word blog post only to have it fail AI detection. And then after doing all that work, having someone refuse to pay.
Then I ran across another one.
Some blogs that accept guest posts are doing the same. They’re running submissions through AI detection tools. Wow.
Some of the blogs judging writers failed their own test…
I found a couple of *very* prominent blogs that say they run guest post submissions through AI detection. I won’t name them because my pockets aren’t deep enough to get my butt sued because of a free newsletter.
Then I went to their blog and grabbed some of their “staff” pieces. Ran them through AI detection. Some of them failed their own test. Oopsie.
There’s a sick kind of hilarity in some guy saying he will use AI detection on your posts if you want to write a guest post on his site, and then having his own writing score out at 45% AI, don’t you think?
Some people are really angry about it…
When I was researching the topic of AI generators, I found one letter posted publicly by a “client” who fired a writer for using AI.
In this case, the writer didn’t deny it. He said look, the article is mostly hand written. Over 80%. But he used AI to help with phrasing some technical details and he’s not trying to hide it. The writing contract didn’t say he couldn’t.
He told the client he should have said so. The client didn’t agree. He said no one hires a writer if they want AI. Back and forth, they argued. Publicly. Yikes.
So who’s right? That’s a good question.
It made me curious. So I set up a test…
The whole mess of people getting accused of using AI made me curious.
Would my writing pass AI detection? I’d hope so, but who knows?
So, I set up a little test. I wrote 100 words on a topic and then asked ChatGPT to write 100 words on the same topic. Then I tested both in half a dozen AI detection tools.
You know what I was hoping, right?
I was really hoping to write about how bad AI detectors suck. A lot of them do suck. One of them tested the ChatGPT copy and said “This is 100% Human. Fantastic!”
I guess if you’re a human trying to pass AI text off as hand written content, it would be fantastic. lol. Except it was an utter fail, because the text was 100% ChatGPT with no human writing at all.
Some of them nailed it. One test said it highlights AI text, and then highlighted the entire ChatGPT piece. That was great! It made me laugh. lol
After my deep dive, I think I see why it happens. I just wonder if the people using AI detection to slam other people see what’s happening.
The irony of AI detection
There’s a real irony to using AI detection tools in the first place. Because AI was trained on us. Literally. They fed it giant swaths of the internet to teach it how to write. Which basically means it learned from us. And now if we write like ourselves, we fail.
It would be like training a robot to write like Stephen King or Margaret Atwood and then when the robot can emulate them, critizing them for writing like a robot.
What a stupid world we’ve built.
I learned something really interesting…
We writers love to talk about voice, don’t we? Trouble is, I’m not convinced anyone really knows what that means. I googled it and some places say it’s the writer’s style.
Problem is, I can write in more than one style. I had to learn when when I was doing ghost writing and writing for corporations. I can read how you write and emulate your writing close enough that most people wouldn’t be able to tell.
Other places say it’s the way your personality comes across on the page. I’m not sure on that, either. I know people who have fascinating personalities and are great to talk to, but put them in front of a keyboard and stuffed shirt comes out.
It’s like they’re still writing a report for teacher.
Is that their personality? Their “voice?” I don’t think it is. Maybe I’m wrong.
Here’s one thing I do know. ChatGPT strings words together the way “most” people do. That’s what happens when you tap into the collective, right?
Turns out the way to “beat” AI detections tools is not to write like everyone else.
I’m going to share my test results on Medium. Probably Sunday or Monday. Hope you’ll watch for that one.
Until then, I’m curious what you think of the whole mess.
On Medium…
Thanks for reading. If you enjoyed this, maybe click the heart to let me know. :)
xo, Linda
Have you ever run your own writing through an AI detector? It's an eye opener.
What a bloody mess!!
Teachers teach off a curriculum, they grade off their own personal experience and take on writing. They WILL correct you into a style even if they don't intend to, it's inevitable. And then the 12th year teacher will try to undo what the 10th grade teacher instilled, because it's not quite right, and guess who's confused?
Of course AI is based on everyone's writing. Computers don't learn (yet) by hacking a library and gorging on every book and then deciding - I think Dicken's style best suits me. They are programmed into a style. That people - including creators of AI - don't see the inevitable pitfalls is mind boggling.
I swear these guys operate like Pot Heads. They're all toking on joints at night, thinking amazing things that will change the world and when they start "creating" it, they're all going WTF, this doesn't seem as good as it did last weekend. No sh*t.
AI driven tractor to plow fields - yeah I get that and can support it. AI developing what we should eat based on "their" knowledge of what's good - no thanks.
Funny how the faster we race into the future, the further we get behind. Gosh, didn't one of the Stoics tell us how to avoid this?!
Cool article!!
I think we're about to see a tsunami of AI-created work on some platforms, while I think the success of others like Substack will be driven(at least in part) by the desire to avoid it.
I also have to wonder how many clients will now be using AI as an excuse to stiff writers. I know not everyone is looking to scam freelancers, but this seems like a really easy way to do it.