Sometimes links in these posts are affiliate links. This means if you click on the link and purchase the book, I will receive an affiliate commission at no extra cost to you.
I woke up this morning to a newsletter in my feed that basically staked the claim that people don’t care — and if they do, they won’t in the future, once they become inured to AI slop — who wrote the piece they’re reading. I can almost believe that, in this age of anti-intellectualism. But Dear Reader, if they don’t care, they should.
Why Am I Telling You This?
I am telling you this because I woke up this morning to yet another dismissal of concerns about AI (“moral panic”) in my feed, and because I have not yet got my jaw up off the floor. And on the way to that dismissal, this particular author on my other platform quipped, “If a student wrote their own essay, they probably engaged with the material, at least insofar as they searched Amazon Books and glanced at a few previews. (I was a bad student. Sue me.)”
I am a former college instructor: Trust me when I say that this did not go down well. That’s not a “bad” student. That’s a dishonest student. The gist of the rest of their argument, if I read it correctly through the red haze of my outrage, was that it doesn’t matter who wrote any given piece published on the internet. What matters is whether the information is trustworthy. They then leap to conclude that it is and that, therefore, the author is ipso facto also trustworthy.
Umm, no. That is begging the question. Maybe begging two.
Because in fact AI is not trustworthy. Not to mention, the very fact of publishing something over your name as if you wrote it when no, you did not, makes you an untrustworthy liar. Not to put too fine a point on it. This is not the only logical flaw in the argument, which may be explained by the fact that, per Pangram, about half the article was genAI run through a humanizer. Fifty-two point four percent, not to put too fine a point on it.
(I would also argue that there’s no way such a writer can assure us that the information is trustworthy if their work ethic today is anything like that described for the student of yesterday. “Looking at a few previews” hardly qualifies as carefully vetting claims before you publish them.)
I’m not citing them here because the point is not to call them out, specifically — it is to make a more general point about who you should and should not rely upon for information. Nor am I interested in getting into a pissing contest with this person. They did not respond in good faith to the last blogger who critiqued their work; it’s not likely they would willingly engage with me. Life is short, and I have no wish to waste the time. If you want to know badly enough, I invite you to Google it.
You Should Not Waste Your Time
Why? Because publishing a piece over one’s name which was not, in fact, written by one is in and of itself a lie and immediately brands a person as untrustworthy. Further, anything produced by an agent like Claude or ChatGPT is already suspect because it cannot think, reason, or make critical judgments based on experience, but can only regurgitate what it has been “trained” on. In sum, anything produced by someone who hides the fact that they don’t do their own work is therefore doubly suspect. Please do not read them. You are putting yourself at risk every time you do.
In the piece I was so incensed about this morning, the OP called checking for AI ‘performative,’ says that in five years nobody will care. They say running work through a checker is ‘lazy.’ A ‘shortcut.’ And they say calling authors out is “entirely beneath consideration,” something a reader only does when they ‘disagree with the work’ or ‘don’t like the writer personally’. (Now you see why I don’t cite them or link to the piece in question?)
Anyway, I disagree with this 100%. I check, I taught my students to check, and I want you to check everything. Everything.
The Why and the How Of It
This is going to be a quick tutorial on why we use shortcuts and how you can use mine. When we’re done, I’m curious how you’ve used them, and whether you have others I forgot to list or didn’t know about, so I hope you’ll comment.
Shortcuts, appropriately used in appropriate circumstances, are a good thing.
OP writes that using a tool like Pangram to vet your sources is lazy because it’s a shortcut to what we really want to know, which is not who wrote it but whether it is sound–and that, they say, in a post-AI world means “you have to measure each piece and each originator,” presumably checking their qualifications (not just reading what they claim as credentials) and checking their sources (but hey: OP offers none) or Googling it for yourself. This is nuts. The whole point of reading is for the writer to teach us what they know, and then they facilitate our engagement with them and each other in the Comments. If they are using AI, they aren’t the one teaching. The agent is. And as I tell my students, if I wanted to know what ChatGPT thought, I’d ask it. Certainly, they aren’t going to be able to engage with us beyond what’s already there on the page–if that. AI-generated discussion posts (comments) as you are well aware from your reading, are breathtakingly vapid and often irrelevant to boot.
Here’s why else that’s nuts.
I can testify, as an instructor with close to 30 years’ experience in the classroom, that checking sources alone is a joy-sucking, time-consuming, pain in the ass that requires access to specialized databases and the professional expertise to evaluate academic papers. If we all had the time and skills for that, we would not be reading stories on Medium or newsletters on Substack or blog posts in WordPress in the first place, now would we?
Sussing out people’s backgrounds is even more exhausting and requires even more technical expertise if they’ve gone to any length to obscure their histories or if they’re not already publicly known (Brown, 2025). If the “person” is themselves a bot, you won’t find anything, no matter how skilled you are or what tech you have access to.
It therefore in no way reasonable to write a piece, say “trust me,” and expect the ordinary layperson not an expert in your field to do their own quality control. In any event, if I could do that I’d have skipped the paid subscription to the platform and the second layer of multiple paid subscriptions on top of that to the individuals who write there.
We need shortcuts. Before we waste time on AI slop, or worse yet make some decision and take some action based on wrong information.
Scanning for AI use is just such a shortcut. But there are others!
My Pet Shortcuts
Here are the tools I use daily to shortcut (bypass) AI drek. First, if the author defends AI use in any way, at any time, that is a major red flag. Second, if they snort derisively at legitimate concerns about AI use, trivialize its consequences, or use false equivalencies to do so. Third, if I see other logical or factual errors. Fourth, if they offer no citations for sources that support their assertions of fact. Fifth, prodigious output not explainable without the use of AI. One of these is enough to screen out (mute, block) a new-to-me author; two or more makes it mandatory in my opinion.
… and Why I Like Them.
The first and second red flags are obvious. They’re telling you, even while they might be denying it, that they do in fact rely on agents like Claude to “write.” Such statements will often read as defensive in tone. (OP’s tone was arrogant, but you know what they say: Sometimes the best defense is a good offense. And they were offensive!)
Third, genAI is famous for logical errors, poor analogies, false equivalences, false dichotomies and so forth–the whole list. OP, for example, trivializes the consequences of AI error by falsely equating searching the net for brownie recipes to looking for solutions to your addiction. We don’t care who wrote the brownie recipe, all we care is that it works, they (or Claude) wrote. But the worst thing that can happen if you land on an AI-generated brownie recipe that’s wrong is that it sucks, you have to throw it out (wasting food), and you’ve wasted time and money trying it. Maybe you’re embarrassed by it. But if you are looking for solutions to your addiction and the AI agent is wrong, you could wind up dead. Or killing someone else, perhaps while baking brownies under the influence. You see something like that in a piece, you have my permission to quit reading. It’s probably AI.
If they offer no citations for their “facts,” it’s probably because they don’t have any, also a hallmark of AI slop. I read somewhere recently (and now I can’t remember where, can’t even find my notes on it) that you need to read three books on a topic — multiple times each — to really know anything about the subject. Or at least, the author advised, three long academic papers, also multiple times each. The blogger who cites nothing is highly unlikely to have done anything close to this. Citing sources is no proof that they did (see Amazon previews, above), but not citing sources is a major red flag.
Prodigious output is always sus, especially if it covers too wide a range of subjects. You simply cannot turn out a long read a day, or several short ones in a day, as this person does without resorting to AI or some other form of plagiarism, including self-plagiarism. Nor can you be an expert in everything. When does this person eat? Sleep? Walk the dog? Read those three books/articles? If this is their only source of income and they’re putting in 50- or 60-hour weeks, maybe. But if not, check to see if there’s some other rational explanation. Think Vincent Van Gogh: Polymaths, people on the spectrum, folks with ADHD, manic/depressives and your standard-issue genius can all churn out massive amounts of high-quality work in a short time — but even then it should come in bursts, with noticeable lulls in between. Even speed-freaks have to rest sometimes. If there’s not some non-AI explanation, check out. Go read somebody who has put time and effort into producing something real for you, who respects and cares about their readers enough to do a good job for you.
But back up a minute, here: Why does it matter? Why do we care who wrote it?
We care, because a student such as OP apparently once was does not know what they are talking about. If they did not do the work when they were in training, they do not have the expertise they are now claiming. And even if they did, if they are using AI now, they are not retaining those skills or staying current in their area of expertise. If we want surgeries done right and the airplanes staying up there in the sky where they belong once students like that are out in the professional world (Bajaj & Sakran, 2026), if we want truly reliable information when we read, then provenance–and the work habits and cognitive abilities behind it–very much does matter.
But again, lastly, verily I say unto you that if Pangram tells you they’re full of shite, they are full of shite. I get nothing for recommending Pangram. If you have something else you like better, go for it (and tell us why in the comments). But keep in mind, I did this for a living: Pangram has the best research data to support it and is the one I trusted for running my online university courses, where the stakes are high and I had to get it right.
The Best Thing You Can Do For Your Quality of Life and Peace of Mind is Unsubscribe From People Like This.
Then go one step further: Mute them.
When I’m done with this and no longer need the original post for reference, I’m un-subscribing. And muting. This is what I recommend you do: Use Pangram wherever it’s freely available. If you depend on the internet for a lot of your information, buy a subscription (they’re dirt cheap and so generous I’m still running on the one I bought last Fall for class). Run it 24/7 in your browser. It could save your life. Remember the red flags and apply that ban-hammer liberally.
Do not let your emotions overrule your critical thinking/judgment. I really wanted to like this person, and so I procrastinated on following up on my first niggling suspicion. I had agreed with a lot of what they said in other pieces. I found some of their advice about writing helpful. They’re funny. They’re smart.
Or at least they seem so? Maybe that’s Claude….
And don’t let the bright lights blind you. Are they rich? Successful? That doesn’t always equate to truth and accuracy! This one is quite successful… but interestingly, is and has been so, it turns out, under many different guises, some of which came with considerable conflict-of-interest baggage. (I, however, would not have known this: Someone else sussed it out and published on it last year. Like I said, this requires skills and access we don’t all possess.)
(And on the subject of conflicts, I have none, at least none that are undisclosed. Or if I do, they’re limited to occasionally getting distracted by what I’m interested in as opposed to what you might like to read about on a given day. Like this post.)
The question used to be, “Is it real, or is it Memorex?” but the stakes are much higher now. The warning to you is that it took me, a college instructor who’s spent nearly 28 years sorting out who wrote what and how for a living, for Pete’s sake, almost a month to catch on. And then I spent a whole morning verifying. Please be smarter than me! Leave these people at the first inkling of trouble.
The poster wrote that in five years nobody will care. My personal and professional opinion is this: I think we will care even more as more and more of the internet is AI-generated and it creeps into more critical areas of our lives than brownie recipes. AI-generated errors can and do cause illness, disability, and death. People who followed its advice are already turning up in area emergency departments (Brown, 2026; Lynch, 2026; Riddle, 2026). Critical readings of the information that we consume can only get more important, as far as I’m concerned, and people like the OP discussed here today cannot be relied on to give you that critical reading.
I have another upcoming article about this (critical reading, that is), in relation to self-diagnosis and self-treatment of anxiety as propounded by the dietary supplement industry. Watch for it! The principles that I will outline apply, whether you have anxiety issues or are taking supplements or not, so I hope you won’t skip it. They’re universally useful scientific literacy skills. Which you may think you already have, and well you might — but you might also learn something new! (I do, every day.) So I hope you won’t skip it.
Meanwhile, what’s been your experience with on-line AI slop? What’s your stance towards AI use in writing? How do you feel about it?
This might be as good a spot as any to say that there will never be a word of AI-generated writing in the Docsplainer, unless it’s a quote or paraphrases from AI-generated work that has been identified by me as such. Like today. That little button on the right of our drafts screen that says “Improve with AI”? Never.
Nevertheless.
If, because there are occasional false positives, you ever scan one of my articles and Pangram or whoever tells you I used AI, please challenge me on it. Publicly. I will be happy to share my entire work process right there in the comments, from original idea to finished product, with complete version histories from the project page down to the individual notes I used and the time/date stamps on my source saves.
As always, if this was interesting or helpful in any way, I’m going to ask you to please… buy me a coffee! This is how I make my living now.
References
Anonymous (2026). Who gives a shit? [or words to that effect — Ed.]. My feed.
Bajaj, S., & Sakran, J. (2026, August 10). What happens when medical students rely on AI – and never develop their own judgment? The Guardian.
Brown, B. K. (2025, November 24). Did Joan Westenberg memoryhole Web3 NFTs? Medium.
Brown, J. (2026, August 5). Consumer Reports warns about AI medical answers. WCVB. https://www.wcvb.com/article/consumer-reports-ai-chatbots-health-information/73354838
Lynch, M. (2026, July 29). Pastor’s Near-Fatal AI Misdiagnosis: Chatbot Healthcare Risks Exposed. The Tech Edvocate. https://www.thetechedvocate.org/a-pastors-near-fatal-ai-misdiagnosis-the-truth-about-chatbot-healthcare/
- N.B.: Pangram says this article is itself 100% AI-generated: Enjoy the unintended irony. And see if you can identify the red flags that made me decide to scan it.
Riddle, K. (2026, March 11). ChatGPT might give you bad medical advice, studies warn. NPR. https://www.npr.org/2026/03/11/nx-s1-5744035/chatgpt-might-give-you-bad-medical-advice-studies-warn
Leave a Reply