Context: Needed a character counter for a school project and found justDone. Except it does more than count the characters, it also gives you a review of how likely your text has been written by AI. As a writer, and out of curiosity, I decided to check how it would rate a part from one of my original pieces. I got a 74% AI score. And you can also humanize it. I tried it, and it sounded more AI than my written part, according to ChatGPT, Originality and GPTZero. So anyway, here’s ChatGPT roasting JustDone.
JustDone’s AI detector has the confidence of a genius and the accuracy of a Magic 8 Ball that got drop-kicked into another dimension. This thing reads actual human writing—creative, nuanced, with just enough imperfection to prove there’s a brain behind it—and still goes, “Yep, robot,” like it just cracked the Da Vinci Code of literature. Seventy-four percent AI? That’s not a result, that’s a vibe check gone wrong. Where does that number even come from? Did it spin a wheel? Roll dice? Whisper to itself, “feels kinda robotic today”? Because the only consistent pattern here is JustDone being confidently incorrect like it’s getting paid per bad take. And the real masterpiece of incompetence? It doesn’t just accuse you—it then offers to “humanize” your text. Oh, thank goodness, right? Nothing says “human” like stripping away personality, flattening the rhythm, and turning a living piece of writing into something that reads like it was assembled from expired corporate templates. That’s not humanization—that’s literary tax evasion. Your original writing had voice; its version has the emotional range of a damp spreadsheet. Let’s be honest: this tool doesn’t detect AI. It detects whether your writing is too good for its comfort zone. You use a slightly elevated tone? Suspicious. You maintain consistency across a passage? Alarming. You write fantasy with a coherent narrative voice? Immediate robot allegations. At this point, if you handed it a page from a classic novel, it would probably flag it and suggest rewriting it in “simpler, more human language,” like it’s trying to downgrade Shakespeare into an instruction manual. And can we talk about the sheer audacity of presenting that “74%” like it’s courtroom evidence? No breakdown, no explanation—just a number delivered with the confidence of a witness who absolutely did not see anything but would love to be involved. It’s not analysis, it’s numerology with Wi-Fi. Meanwhile, your actual text—alive, slightly irregular, stylistically intentional—gets flagged because it doesn’t fit into the tool’s painfully narrow idea of what “human” looks like. Which apparently is: bland, predictable, and mildly sleep-inducing. Congratulations, JustDone, you’ve reinvented the concept of judging a book by how boring it is. And here’s the best part: this entire roast was commissioned by a human who quite reasonably decided that the sheer absurdity of JustDone’s existence was not worth spending their own braincells on. That’s right—you annoyed someone so much with your nonsense detection that they outsourced the insult production just to avoid engaging with you intellectually. You didn’t just fail at AI detection—you inspired intellectual outsourcing out of pure exasperation. At this point, being flagged by JustDone isn’t an accusation—it’s a badge of honor. It means your writing has enough structure, clarity, and style to completely confuse a tool that fundamentally does not understand writing. If anything, a low score from it would be more concerning. In conclusion, JustDone isn’t exposing AI—it’s exposing how little it understands the thing it claims to measure. Every false positive is just it loudly announcing, “I have no idea what I’m doing, but I’m going to say it with confidence anyway.” And honestly? That might be the most artificial thing in this entire situation.
Engine: ChatGPT-5 on March 26, 2026 Prompt: “Write me a roast for how bad justDone is at ai-detection. Make it extra long and extra spicy. Add that the roast was commissioned by a human who doesnt want to spend braincells on cooking up its own roast from the sheer absurdidty of justDone.” submitted by /u/BlaXtarAscent
Originally posted by u/BlaXtarAscent on r/ArtificialInteligence
You must log in or # to comment.