Listen to this ►

The “AI Detector” Grift Game: Why They Have Dumb Down Human Writing So As To Not Get “AI Flagged.

Be sure to run this through an “AI Detector “ as it will be flagged “100% AI”.

I want you to imagine a scenario for a second and really just try to put yourself in this person’s shoes. You’re a college senior. You have just spent the last three months meticulously researching, outlining, and crafting your final capstone essay. Or maybe you’re like a mid-level manager at a logistics firm, and you’ve spent weeks agonizing over this crucial market analysis report for the executive board. You poured over the primary sources. You refined every single sentence. You checked your transitions. You polished the vocabulary. You made absolutely sure the logical structure was completely flawless. And there’s that singular moment of relief… You hit submit, you lean back in your chair, and you feel that specific rush of pride that only comes from a job genuinely well done. But the relief just doesn’t last. Because almost instantly, a piece of black box software, a program you have never seen… A program whose internal rules you are literally not allowed to know. It scans your months of hard work and slaps a giant neon red completely fake label across the top of it.

It’s devastating. It calls you a fraud to your professor or to your boss. And in a matter of seconds, your grade is ruined. Your professional reputation is in tatters. And worst of all, there is no human being you can appeal to. The machine has spoken. It is a genuinely terrifying position to be in. The psychological toll of that just being called a liar by an algorithm and having no way to defend yourself. And I think we need to be clear right from the start here, this is not some dystopian science fiction hypothetical. It’s happening right now, this very morning, in high schools, universities, and corporate HR departments all over the world.

The entire industry of text based “AI Detectors” is a pure grift machine built on arrogance, bad science, and the willfull stigmatizing of millions of real human writers. These tools do not work. They never worked. They will never work in any reliable way that matters. And the people selling them know it.

It is the height of irony that AI models stole the written works of you and me—and are now the judge and jury of whether humans wrote the very words they stole from us. Only to be metered back with the human content dumbed down and subservient to AI output. This is how bad it has become, and it will get worse if you and I dont stand up to these freaks.

I wrote a while ago about the use of Ai in any work of text in detail. I predicted some of these “AI Detector” companies would begin and it has. they are desperate for business through any means necessary.

You Have 5000 Days: Navigating the End of Work As We Know It. Part 10: Everyone Is Doing It*.

And:

Rethinking AI Testing: Beyond The Numbers.

Also the game of fingerprinting and watermarking AI output.

How to Clean Anthropic’s Fingerprinting And Watermarks With This New App.

This article is sponsored by Read Multiplex Members who subscribe here to support my work: Link: https://readmultiplex.com/join-us-become-a-member/

It is also sponsored by many who have donated a “Cup of Coffee”. If you like this, help support my work: Link: https://ko-fi.com/brianroemmele

Listen to the companion podcast: https://rss.com/podcasts/readmultiplex-com-podcast/3093999

Problem, Reaction, Solution

Of course people are using AI, and calculators, and cars instead of walking and electric lights instead of candels. It is up to the author to disclose if they used such tools but these companies claim to solve a problem that does not exist in the way they pretend. The only questions that actually matter when you read something are simple: Is this passage worth reading? Is it thoughtful about the subject? Did I learn anything?

An average human can answer those questions without some black box software that spits out a fake percentage and then ruins a student’s semester or a writer’s reputation. Detectors solve nothing. They create a new form of digital scarlet letter based on statistical noise.

There is no way to prove “provenance” of text once large language models have hovered up nearly everything ever written. The training data is mostly the sewage of the internet. Reddit posts, forum rants, low effort blogs, the endless “yo bro” casual style that these systems now treat as the gold standard of “human.” Formal, careful, structured, or older writing gets treated as suspicious.

That is not just the bug. No—that is the product.

Look at what happens when you feed them real human text from before any of this existed. Passages from the Bible regularly score as heavily AI generated. The Koran gets the same treatment. The US Constitution has been flagged over and over. Peak human writing from the 1950s, Shakespeare, formal academic papers written decades before ChatGPT, all of it gets scanned and marked “AI.” The other way the grift works is they use text associations from the past show up like it is not AI. This is to help sell the worthless system to schools that test it on past work.

Famously ZeroGPT has scored Genesis at something like 88 percent AI. Other detectors do the same to classic literature and religious texts because those texts are formal, repetitive in structure, and low in the kind of random “burstiness” the detectors were trained to expect from modern internet sludge. This is not occasional error., but a systemic issue of the entire industry.

Oh they have trained some of these models to know the commn texts so as not to get caught in claiming The Bible is “AI Generated”. But the exceptions list grows so as to continue to farm millions of dollars from schools and universities to be “sure” they have an AI solution. Sometimes reenforeced by state laws. A perfect grift these companies are.

Multiple independent studies have shown the same ugly picture. Weber-Wulff and colleagues tested 14 detectors including Turnitin and various free tools. Every single one scored below 80 percent accuracy. Most leaned toward calling AI text “human” while still producing enough false positives to destroy trust. Accuracy got worse when text was paraphrased or translated. The tools were “neither accurate nor reliable.”

Now look at the University of Florida research (https://www.semanticscholar.org/paper/AI-Wrote-My-Paper-and-All-I-Got-was-This-False-the-Layton-Medeiros/32de00afd2ab52521d9257b428ea3efafd827200), because this one really puts the nail in the coffin. Patrick Traynor, professor and interim chair of the UF Department of Computer & Information Science & Engineering, along with Seth Layton, Bernardo Madeiros and Kevin Butler, presented a paper at the 2026 IEEE Symposium on Security and Privacy titled “AI Wrote My Paper and All I Got Was This False Negative: Measuring the Efficacy of Commercial AI Text Detectors.”

They did something clever and brutal. They took abuot 6,000 real research papers that had been submitted to top-tier security conferences before ChatGPT even existed. Then had large language models generate AI clones of those exact same papers. Both the original human papers and the AI clones were run through the five most popular commercial detectors on the market.

The results were a disaster. False positive rates ranged from 0.05 percent all the way up to 68.6 percent. False negative rates swung from 0.3 percent to 99.6 percent. That upper number means these tools basically missed almost every single AI-generated paper.

Read that again. And again.

The detectors collapsed. They became largely useless, unable to tell the difference anymore.. Traynor’s conclusion was clear: these tools are not reliable or robust enough to use to measure the problem and likely never will be. Commercially available AI-generated text detectors are “poorly suited for deployment in academic or high-stakes contexts.” People’s careers are on the line, he said. And yet universities and journals keep buying the fantasy.

Which brings us to the worst of the worst: Pantagram (purposeful, but really ironic, misspelling) Labs and its “I know better than you-type) founder Max Spero. This is peak arrogance wrapped in Stanford pedigree. Spero and his co-founder Bradley Emi both came out of Stanford with fancy AI degrees. Spero did time at Google (obviously). Emi worked at Absci.

They sell themselves as the serious, high-academic solution. Spero even styles himself the internet’s “slop janitor,” publicly scanning journalists and writers and calling them out by name when his tool flags them.

Pantagram claims absurdly low false positive rates, one in ten thousand, sometimes even lower in their marketing. They position the product as the reliable academic-grade tool that universities and conferences should trust. And some of them do. NeurIPS 2026 sadly used Pantagram to screen position papers and desk-rejected 178 of them with no appeal process based on the scores. That is not a tool. That is a gatekeeping weapon. The fine folks at SubStack, (I am talking to them about this now) have been scammed to use it as a “writer’s tool”.

How The “Ai Detector” Grift Works—The Gatekeepers Of Knowlage

Here is the real game. These detectors are being sold to the university machine so it can protect its monopoly on knowledge. The academy has always controlled what counts as legitimate knowledge production. Now that AI threatens to let anyone generate competent text, the system reaches for tools that can taint anything not produced inside its approved channels as “AI Slop.” Flag the outsider writing. Flag the independent researcher. Flag the student who writes too cleanly or too formally. Keep the credentialed pipeline pure. Pantagram and its competitors sell that service while waving their elite resumes.

Read this again and understand: AI threatens to let anyone generate competent text, the system reaches for tools that can taint anything not produced inside its approved channels as “AI Slop.”

The irony is extreme and almost too perfect. These same companies and the systems that buy their products are themselves swimming in AI. Marketing copy, technical reports, blog posts, even internal tools in this industry were vibe coded with the very technology they claim to police. Max Spero’s own company materials and the polished academic framing they push sit right next to the detectors that would almost certainly flag large parts of the older human literature they claim to protect. The people selling the purity test are using the impure tools. The university that wants to declare outside knowledge “slop” is using a product built by the same AI ecosystem it pretends to fear.

Even OpenAI Gave Up With “AI Detector” Grift

OpenAI itself tried building a classifier and then shut it down because the accuracy was garbage. They admitted it labeled human text as AI 9 percent of the time while only catching a fraction of actual AI text. The company that built the models most people worry about could not make a detector that worked. That should have been the end of the conversation. Instead a whole industry of GPTZero, Originality . ai, ZeroGPT, Pantagram and the rest just kept selling the fantasy.

The companies and their shills are peak arrogance. They sell fear. They sell the idea that without their product the academy or the workplace will collapse into AI sludge. Then many of them turn around and offer “humanizer” tools that take AI text and rewrite it until the same detectors they sell say it is human. The cycle is perfect grift. Create the panic. Sell the detector. Sell the way around the detector. Repeat.

They shame real writers. Students who write carefully get flagged. Professionals who produce formal reports get flagged. Anyone who spent years learning to write with clarity and structure gets told their voice looks like a machine. The stigma is real. Careers and grades and trust get damaged on the basis of a probability score that the science shows is unreliable. And there is almost never a clean way for the accused to prove innocence because the detector companies treat their methods like trade secrets while waving around marketing claims of 99 percent accuracy that independent tests never support.

This is a massive issue as we move deeper into the AI age. We have an arrogant founder using his “AI Dector” to publicly shame people in hopes to silence them. If this sounds like 1984, it is because it is—like 1984. The fear of sounding too knowable or too practiced in your prose will land you into a thought control and ultimately—state sponsored dumbing down of human copra to satisfy the “AI Detectors” that are trained to think we all “yo bro” write with our thumbs on Reddit.

This danger is real and I have seen it impact many great writers that now fear to use even a spelling correcter so as not to be “fagged as AI”. We will not be shamed into not writing our thoughts— AI ASSITED OR NOT. We will not quietly go silent into the dark night.

Oh It Gets Worse

As the models improve the problem only gets worse. Better models produce text that looks more like the best human writing in the training data. Detectors trained on older patterns fall further behind. Humanizers and simple prompting tricks already break most of them. The arms race favors the generators, not the detectors. Papers keep showing the same pattern: high claims on controlled benchmarks, collapse under real conditions, domain shift, light editing, or newer models. And the entire desperate scam will grow wider and wider. As they desprealy try to get funding for the grift. Well I am taking measures to be sure VCs know what this is before they invest and help the ones that already made the mistake.

There is no technical fix coming that turns these tools into reliable proof of authorship. Text is not a fingerprint. Ocne these models have seen enough of the world’s writing, statistical detectors are chasing shadows. They measure predictability and style patterns that humans also produce, especially careful or traditional humans. The more the models improve, the more the overlap grows.

Moving—Forward

Reading this— you must remember that someplace out there are great writers in High School, Universities and on SubStack and blogs that are getting the Scarlett Letter from usually wrong “AI Detectors” as a new witch hunt and purity test. I have spoke to dozens of kids in the summer that got caught up in this and I will tell you lawsuits are on the way and I am a witness to some of these cases, so I am not guessing. These young lives have been permanently injured by these grifters. If you agree,l you must speak out. We must not accept this on any level. We humans will not dumb down our writings to “Yo Bro” styles so some “AI Detector” gives them a “pass”.

The honest position is simple: These detectors do not work for high stakes decisions. They don’t work on low stakes detection. Using them to implicate people is—reckless at best and criminal at worse.

The real evaluation of any piece of writing remains the human one, you already own this detector for FREE, is the writing:

is it good,

is it useful,

does it show thought.

Everything else is theater sold by people who profit from the fear they helped create, especially the ones with the fancy Stanford resumes who sell the university its new purity test while the whole system quietly uses the same technology it pretends to hate.

I started this article by talking about how these detectors assume the messy, slang-filled sludge of the Internet is the baseline for human writing. I call it the sewage standard. So think about the downstream effects of that over the next 10 years. If the only way to mathematically prove you are human to a machine is to write poorly, to use slang, to be deliberately disorganized and erratic in your sentence structure, are we about to see an entire generation of brilliant students deliberately dumbing down their writing just to avoid being accused of cheating by an algorithm?

It is a chilling incentive structure. You’re essentially training a generation to hide their intellect. If excellence gets you flagged for an academic integrity violation and mediocrity gets you a passing grade, what does that do to the future of human literacy? We might be entering an era where writing beautifully is considered a crime. And that is something we all need to think very, very critically about the next time someone suggests installing one of these tools.

A vital warning.

The industry is a fools errand dressed up as solution. It will keep getting more useless as the underlying technology advances. The only people still pretending otherwise are the ones collecting subscription fees and the institutions desperate to keep their monopoly on what counts as real knowledge.

Resist these fools today—while you can. Not because it is the right thing to do, but the only thing we have as humanity.

To continue this vital work documenting, analyzing, and sharing these hard-won lessons before we launch humanity’s greatest leap: I need your support. Independent research like this relies entirely on readers who believe in preparing wisely for our multi-planetary future. If this has ignited your imagination about what is possible, please consider donating at buy me a Coffee or becoming a member. Value for value you recieved here.

Every contribution helps sustain deeper fieldwork, upcoming articles, and the broader mission of translating my work to practical applications. Ain ‘t no large AI company supporting me, but you are, even if you just read this far. For this, I thank you.

Stay aware and stay curious,

🔐 Start: Exclusive Member-Only Content.

Membership status:

🔐 End: Exclusive Member-Only Content.

~—~

~—~

~—~