As a writer, I run into a lot of scams. I typically spend a lot of time checking job boards for openings, and I’ve landed a few one-off jobs and a few longer-term clients on social media like Reddit and LinkedIn. Recently, however, I’ve noticed a trend that some “enterprising” companies have come up with. In this article, I want to look at this AI-checker scam, why it’s something freelance writers should worry about (especially for low-paid jobs), and why anyone who tells you, as a writer, that your content needs to pass some amorphous AI-checker test is just trying to steal your work from you for free.
What’s an AI Checker?
According to some scam artist companies selling services out there, an AI checker is supposed to determine whether something is AI-written or not. That’s a pretty vital function if you’re a company that wants your content done by a real human being. Writing (such as mine) comes from my head and aims to connect with you, the reader. All good content offers the reader something. Yet AI writing tends to have a lot of fluff, a healthy helping of buzzwords, but not a whole lot of actual content. Well, if you just shoot it a prompt and let it go, that is.
Now, here’s the kicker. None of these so-called AI checkers actually work. Absolutely none of them. They are there to give a result that says your work is AI-generated when it isn’t. A case in point is this piece of content you’re reading now. Every word from the intro to here has been written by hand, from my head onto this page. Yet, this “AI-Checker” claims that 21% of my text is AI written. As someone who profoundly understands AI models, I can safely tell you that these companies are lying to you. But the game here is not to tell you whether your content is AI-written or not. Oh no. This goes a LOT deeper.
My Brush With AI Checking
I’ve been actively emailing new clients and searching for jobs over the last few months since I have some bandwidth opened up to work on new content. One of the companies I pitched asked me to do a sample for them. Of course, being someone who likes to prove myself, I took an hour of my time, did the requisite keyword research, plotted the content out, did headers and meta descriptions, and crafted a 1200-word article that could easily rank on a search engine page. I fired it off without thinking much about it. Then I got the reply, and it left me confused and angry.
You see, Constant Reader, this company ran my content through their AI checker. Yes, one of those scam tools this entire article is about. And the scam tool gave them a reading of “around 29.73%” for AI content. Naturally, I asked them to please let me know how they came up with this number. As a scientifically-minded person, I would always like to know how these checkers work. I’d like to know why it gave my 100% handwritten content a 29.73% likelihood of AI generation. How did it derive that with such accuracy? I haven’t received a reply from the company since, and I think we both know why that is.
Something For Nothing (And Clicks For Free)
Aside from my personal experience. I’ve come across a large volume of posts on Reddit about these companies that require their writers to hit 0% on an AI checker. Some of the poor guys and gals working on this content have been pulling out their hair to hit this number, which doesn’t make any sense. Why would a company ask that a writer hit 0% on a fake AI checker to verify that they’re not using an AI tool to write? And what’s wrong with using an AI tool to help write your content anyway? From a professional perspective, AI is a tool, no more and no less. However, companies forcing their writers to go through this process are not concerned about the practical parts of your content. They don’t care that it’s good content or that it engages your audience. No. They just want “proof” so they don’t have to pay you.
See, many of the companies that offer these jobs to writers have a paltry pay schedule attached to each piece of content. Usually, they pay between one and three cents a word for a five-hundred-word piece of content. Yet, they’re offering this content to clients that are willing to pay a lot more. In most cases, if you’re paying someone a pittance to do good content for you, you’ll still be making a boatload of money. But these businesses are not satisfied. Instead, they prefer to take the “AI-generated” content, refuse to pay for it since it’s AI-generated according to their “checker,” and then sell it on to the client for the full markup. What a scam.
How Would An AI Checker Even Work?
I have a pretty good understanding of how AI systems work, including how they generate ideas and even speak to you. GPT and its contemporaries are really an innovation in technology. So, in that case, how would I do it if I were to develop an AI checker? Well, theoretically, AI follows some recognizable patterns. AI-written content tends to recycle some phrasings throughout its writing. If you’re using AI to write, it’s also apparent in how it approaches the writing process. AI engines are great at pattern recognition but can’t think like humans do.
To detect AI writing, I would check for the presence of these patterns at predictable locations. For example, the conclusion of a piece that AI does summarizes the entire article and doesn’t do Calls-to-Action (CTAs). I’d also have a list of recurring phrases the AI likes to use. Different AIs, for example, use different phrases, so you can tell by which choice of recycled phrasing it uses, which engine is being used, and for how much of the content. That said, even knowing all of these things, it’s impossible to tell the difference between human and AI writing aside from intent. Humanity is found in intent and engagement, which AI can’t do just yet.
A More Sinister Situation
To wrap up this foray into AI-checking scams, I’d like to appeal to companies (legitimate ones) that use them to check incoming content. Using these things breaks the trust between you and your freelance writer. Trust is like a China plate — sure, you can fix it, but it’ll never be the same again. A professional might use AI to help write your content. At the very least, writers like me use AI to help us outline a piece before starting in some cases. It points us in a direction to find the engaging stuff about a topic. If you want 100% human-written text, we can also do that, although it takes us longer and will cost you more as a result. If, however, you’re using these AI checkers to determine whether a writer is a good one or not, then you probably don’t understand AI or writing enough to understand what’s good writing and what isn’t. And nitpicking over it will just make writers leave you as a “difficult” client.
For the writers who read this, I advise you to quit working with people who put you through this. The ones who know that the AI checker is a scam just want you to work for free. And hours you’re not billing for are wasted hours for you. Any content agency (or mill) that asks you to make your writing 0% AI is trying to chisel you out of your work, and you can (and should) do better. Eventually, when these scam artists realize they’re not getting anyone to work for them for free anymore, they’ll probably start offering the pittance they were offering before. Subjecting yourself to this kind of torture for a few cents a word is just not worth it.
Hi, I’m Jason, and I do freelance writing, editing, and professional ghostwriting. I’ve done thought-leadership articles and LinkedIn posts that have ranked on SEO engines and placed my clients as experts in their field. I’ve done in-depth articles on AI and deep dives into the industry, and I am currently working on a course about writing with AI for professionals. I’m also open to new work. If you’ve got something you’d like me to do, you can reach me at email@example.com. You can also subscribe to my Medium page for more articles like this. And remember, if you’re a writer, take yourself seriously. Know the worth of your time, and charge your clients accordingly.