# Scalzi – Another Reason Not to Use "AI" for Your Writing > Source: > Published: 2026-08-01 13:06:10+00:00 And make sure your lawyer doesn’t use AI when they do lawyer things for you. # Another Reason Not to Use “AI” For Your Writing Posted on [July 31, 2026](https://whatever.scalzi.com/2026/07/31/another-reason-not-to-use-ai-for-your-writing/) Posted by [John Scalzi](https://whatever.scalzi.com/author/scalzi/) [ 43 Comments](https://whatever.scalzi.com/2026/07/31/another-reason-not-to-use-ai-for-your-writing/#comments) [ Back in February I did a long post about “AI” and why I wouldn’t use it for my own writing,](https://whatever.scalzi.com/2026/02/14/10-thoughts-on-ai-february-2026-edition/) and one of the reasons I gave was “the fact that ‘AI’-generated text is not copyrightable and I don’t want any issues of ownership clouding my work.” Well guess what? [A hot and buzzy debut novel has been pulled from publishing and film/TV consideration over concerns that “AI” was used to make it](https://deadline.com/2026/07/ai-questions-debut-novel-book-film-buzz-1237013254/). Here is one of the relevant bits from the article: “*Basically, questions were raised over whether the author used AI as a tool in the writing of the book. The problem: publishers and movie companies may not be able to register copyright and come away with a clear title chain if there are certain AI elements in the mix. That’s because AI-aided copy is an amalgamation of many other copyrighted works.* “*From what I’m hearing, this will be an important teaching moment for writers with aspirations to become serious bankable authors. The lesson: stay the hell away from AI.” * Am I feeling smug about having called this six months out? Yes, a little, but mostly not. I don’t deserve credit for noting this would be an issue because it was already a clear legal point: “AI”-generated work is not copyrightable, so if you have any “AI”-generated work in your material you open the door to legal issues. But also, the concern here is slightly different: The issue appears not to be that “AI” material isn’t copyrightable, the issue is that “AI” is trained on copyrighted material, so prose generated from it could be contested by the original copyright holders. Now, from a practical point of view this could be a stretch to prove unless the “AI” farts out entire paragraphs of prose unaltered from its training material ([which, to be clear, it can do and has done](https://arstechnica.com/ai/2026/02/ais-can-generate-near-verbatim-copies-of-novels-from-training-data/)), and those end up in the finished book. But if you’re a lawyer for a publisher or a film/TV production house, even the hint of possible legal entanglement is enough to raise the shields. It’s already in contracts for both film/TV and for publishing houses that the author attests their work is wholly original, so, again, any hint that anything is other than that brings things to a screeching halt. As we have seen. This is a situation that is now always on the table for writers, particularly newer writers with no track record and nothing available for publishers and producers to judge their work against. How *do *newer writers fight against an allegation of using “AI” to write their work? Publishers and producers are now trying to vet new work with “‘AI’ checkers,” [but the problem with those is that they are also “AI,” and they are unreliable as fuck](https://lawlibguides.sandiego.edu/c.php?g=1443311&p=10721367), so we’re back to where we started. It’s a bad situation. Not great for any writer these days! But again, especially so for newer writers without a backlog of pre-“AI” material at their disposal that can, at least, show the writer was already writing without assistance for years. And in fact, as I note above in the attached Bluesky post, I think that for the reasonably near future, it’s possible that older works — ones written prior to, say, November 2022, when ChatGPT made its public debut — will become more desirable for film/TV optioning and adaptation, because they were clearly written without “AI” assistance, and there is a clear chain of title when it comes to copyright. Why bother with new stuff you can’t be confident about? Or, if you *are *going to bother with new stuff, you’re going to go to the new stuff from the writers who have lots of old stuff (i.e., could obviously write without “AI” assistance). If they have a long-standing and public “No-‘AI'” stance, so much the better. Which is great for writers like, uhhhhhh, *me*, and don’t think I haven’t already suggested so to my film/TV reps. But, again, it’s not great for newer writers. Which is unfair. It’s not fair that newer writers will have to prove (to a larger extent than more established writers) that they can actually *write*, and that what they’ve written is actually their own work, not the result of a prompt. It’s especially unfair when the makers of the tools writers use to write — I’m looking at you, Microsoft and Google — are frantically stuffing “AI” tools into the programs writers rely on. No, Microsoft, I don’t fucking want Copilot to “help me write” in Word, and no, Google, I don’t fucking want Gemini to “help me write” in Google Docs. The very fact these “AI”s are hovering around my writing at *all *is bad enough from a provenance standpoint. These “AI” tools are making it harder for all writers in this regard, not easier. I don’t think at this point in time anyone is ever going to accuse me of using “AI” to write anything; I’ve been loud about it long enough, and have a long enough publishing history, that it’s clear that I don’t and won’t. But how will a newer (or lesser-known and less publicly loud) writer prove their writing is their own? Will they have to use versioning to show progression of the manuscript? Email chapters to their editors as they go along? Revert to typewriters and handwritten drafts? At some point it’s entirely possible that publishers (or the publishers’ lawyers) will require writers to “show their work” with regard to the novels they submit — be able to present concrete evidence that everything they write came out of their brain, and not a prompt. In any event, the *Deadline *article is correct: If you’re a writer who hopes to get your work into film and TV, or grab one of those big publishing deals, stay the hell away from “AI.” Don’t incorporate it into your writing process; every part of the process needs to be demonstrably free of any “AI” input because “AI” training data is full of other people’s copyrights. It didn’t have to be this way — “AI” companies could have just as easily trained their LLMs on public domain or licensed material instead of grabbing pirated works and letting God sort it out — but what they should have done and what they did do are two different things, and we have to live in the now. And in the now, “AI” inputs are inherently untrustable, from a copyright perspective. Don’t incorporate “AI” into any part of your *publishing *process, either: Don’t use “AI” editing, don’t use “AI” art, don’t use “AI” translation, don’t use “AI” *anything *that calls into dispute whether your work is actually yours. Using “AI” cover art, for example, will immediately call into question what else in the book is “AI.” Every other aspect of the book is implicitly tarred with the same “AI” brush. When I say that I have it in my contracts that I require every aspect of my book production to be done by humans, it’s not just because I want to honor human work and input. It’s also because I’m protecting my own reputation by doing so. Just… don’t use “AI,” okay? Take comfort in the fact that millennia of writers and authors and storytellers got along just fine without it and you can too. None of them were so special that you can’t do what they did. And this way, if a lucky break comes your way and a publisher or film/tv studio wants to throw literal millions of dollars at you, you will not have given the legal department an easy way to back out… and you will not have a reputation (deserving or otherwise) for not being able to write. Or a reputation for wasting everybody’s time. *That *reputation will follow you, for sure. — JS ### 43 Comments on “Another Reason Not to Use “AI” For Your Writing” - - Maybe I’m taking things too literally (yay AuDHD), or because my Wife is a PhD computer scientist and I know the difference between an LLM and machine learning, I have a real hard time with a lot of these “Don’t use AI AT ALL” arguments. I asked an editor once who had a very militant anti-AI stance if that meant that if using a spelling or grammar checker counted in that because, that’s technically AI, too, albeit machine learning. Or, say, if I used something like Dragon Dictation. A lot of those kind of things use machine learning to make things better and easier. The other thing I worry about with this anti-AI fervor, and that not a lot of more public authors seem to think about, is that it’s going to eventually get used as a way for certain sectors to tank marginalized authors. It’s been happening a lot in fanfic, where someone doesn’t like someone’s work for whatever reason, then accuses them of AI and let’s the internet do it’s work. If you’re a new/indie author, that accusation will stick no matter how well you can prove it. It shouldn’t just be on an author to not use AI, but also on other writer’s to not just throw around AI accusations because they don’t like someone’s writing, or think it’s “bad” by their own (arbitrary) standards. As you point out, no one can truly tell if a writing is AI, nor do any of the programs that can “detect” AI in a piece of writng really work. Don’t get me wrong here– I truly do think that LLM’s suck, are complete crap, and I’m completely down for the AI bubble to burst, but the anit-AI fervor is causing a lot of harm to new and indie authors just because their writing is new or different. - I tend to side with @LJ here. I’m not an author of books, but I do write several blogs and I run a couple of side businesses in my various skill sets. I use AI as an *aide*not to do my writing or my code or anything else.(As a side note, my day job is for a fortune 500 company that is insisting we use AI – to the point that our annual performance reviews are contingent on showing that we’ve incorporated AI into our workflow to make our jobs easier/faster/more productive. So learning to use AI responsibly has been a big part of my process.) But at this point I don’t know that it matters anyway. I wrote a post for my Substack a few days ago and on a whim I used the “scan for AI text” feature that Substack provides – and it told me that my post was written by AI. It was not. It wasn’t even edited by AI. It was entirely written by me. And I’ve seen other writers who I follow on FB and elsewhere being accused of using AI when I know they don’t. So … *shrug*… does it really matter at this point when people who write everything themselves are being accused of wholesale LLM/AI use and there’s no way to prove otherwise? - Stephen Colbert had an interesting view of AI. He was asked in a Slate interview about what a new event was to be a new BCE/CE. One example he gave was the nuclear age with the low radiation steel but then he gave the more recent one of pre/post AI and how things will be looked at differently. It was a great thought. [https://youtu.be/36Ik9CgC1m8?si=EKaO86NKZg-B33sM&t=2280](https://youtu.be/36Ik9CgC1m8?si=EKaO86NKZg-B33sM&t=2280) - @LJ Spencer, your first paragraph is definitely one of my irritation points with the AI techbros who purposefully took all the various machine learning and similar work that had already been done – in many cases, due to the specific requirements and narrow functionality, done very well! – and got it lumped in with their useless LLM garbage under the ‘AI’ label. That does a major disservice to a lot of prior work in the computing space, and my feeling is getting those items separated back out from the ’AI’ label could be a good thing. - How does one prove a negative as a new writer (no AI use for any purpose)? Avoid MS Word and Google Docs and use, for instance, Scrivner? - Ah, right, cool. But I still think we need more discussion about the way some folks are really quick to say someone’s writing is AI and no matter how much a writer can prove otherwise, it’ll stick. Like @Kara, I’ve seen a lot of commenters come on to people’s posts and say, “yeah, this is AI! I can tell!” (Reddit is the worst for this.) And I’ve seen so many new and young writers who are terrified to “sound like AI” and making lists of punctuation — punctuation! — that they feel they can’t use because they’ll be accused of using AI. Humans are going to human, you know? - Yeah, I hope so, too, but, as folks say, nuance on the internet is dead. - that last was in reply to @Mithramuse… dang thumbs… - As a new author myself (first fiction book published July, 2026) I worry about whether people will think I use AI to write. I don’t, but I’d be hard-pressed to PROVE it. OTOH, I have nonfiction pubs going back to the 1980s, including a sole author publication in J Chem Phys in 1982. Wonder if that would cut ice with anyone? - I have not written a book (yet), but I have been blogging for decades. I have been using Grammarly for just about as long (for correcting missing commas, misspellings, etc.). I never really thought of it as AI, but I guess it is. Is it wrong for me to use it as such especially if/when i get around to writing a book? - One area where generative “AIs” are useful in the creative process is assistive technologies like speech to text. ML products like Dragon (is that still around?) probably can’t keep up at all. I’ve seen people using them and they’re game-changers for people with disabilities that make it hard for them to type. Watching someone speak into their phone and the phone just “knows” where the punctuation goes, what gets capitalize, and even getting paragraph breaks right, is just amazing. That is of course an area where the *training data*is highly likely to have been stolen, but the output*for the end-user*is entirely their own work. - I use Grammarly mostly for misspellings and comma placement (though I ignore half of its suggestions). Sometimes I accept word suggestions, but most of the time I find that they flatten out my language. I don’t think it would count as AI if you only used it for punctuation and spelling, or word replacements that you could find in any thesaurus. - Thank you. I feel better about using it now. - For proving no generative AI, is it helpful to use word processor’s “save my history” functions? Word, Google Docs, and LibreOffice all have some version of this, which could be useful for showing changes over time. Even if it doesn’t save keystroke by keystroke, showing the types of edits and changes made over time could help prove to editors, publishers, or legal departments that you were making the kinds of changes that one would expect from a human-written work. - While I spent much of my career as a pilot, at various times I made my living as a tech writer and/or editor of aviation magazines. This was pre-AI, so I never felt its lack (and, to the best of my knowledge, have never used it since). For me, the height of technology at the time (and a very welcome advance) was the IBM “Selectric”typewriter, because it that nifty white film ribbon that made the backspace key the equivalent of “delete” on a computer or word processor. Today’s NYT and NPR news report that an AI model has escaped its “sandbox”testing environment and hacked other computer systems without being detected. It’s likely that there have been, and will be, many other such incidents. Meanwhile, we have the issues of data centers’ massive requirements for land, energy, cooling, and memory; the increasing concentration of AI power into relatively few entities and corporations; and even, Thlord forbid, plans to put “millions of data center satellites into orbit.” So much for my retirement hobby of astrophotography… Not to detract (or distracts) for a moment from John’s, and other writers’, concerns about AI’s encroachment into the creative process. But it seems to me, compared to all of AI’s other existential threats to us as humans, that these may fall into the category of “rearranging the deck chairs on the Titanic.” And FWIW—the human brain, running its “wetware” for actual human intelligences, consumes all of about 20 watts of power. About like a couple of nightlight bulbs before they were replaced by LEDs. - …and I should add that the typographical and grammatical errors in my post above prove that no AI was used in writing it! - “For proving no generative AI, is it helpful to use word processor’s “save my history” functions?” Interesting question. I keep most of my docs on MS OneDrive, and it has a “Version History” function on the web interface, that you can use to go back to very old versions. (For instance, I just used it to look at a doc I started work on in 2016.) I don’t know how secure those saved versions (and their timestamps) are. - There’s a lot of hype and anthropomorphism in the reporting of the recent “sandbox” breach, enough that it’s hard for me to be certain of what exactly happened. But in broad strokes it seems to be basically the same sort of incident as the “Morris worm” of the 1980s, where someone wrote a not-fully-vetted piece of code, deployed it in an environment where it had too many privileges, and it managed to breach other insecure enviroments. (Wikipedia has a decent summary of the incident at [https://en.wikipedia.org/wiki/Morris_worm](https://en.wikipedia.org/wiki/Morris_worm))The main difference between then and now is that we prosecuted the person who deployed the Morris worm, and pushed those running the environments it exploited to better secure their environments. But now we reward both the exploiters and exploitees with venture funds, IPOs, and lucrative contracts. Not only that, much of the tech industry is going full speed ahead encouraging the further propagation of inherently insecure “AI”-generated code in insecure environments. That’s a choice, and one where I see responsibility falling squarely on the people, companies, and governments that enable and promote this sort of exploitation. If we are indeed in a Titanic-like situation, we can get out of it not by wringing our hands about deck chairs, but taking action on who we let control the ship, and what we let them get away with. - Amen. What we call AI (which is really just fancy pattern matching) is horrible in every way: It’s blatant theft backed by climate destruction and increasing wealth inequality. We need to burn it down. Does AI have its uses? Yes, against carefully curated data. The LLMs are worse than useless, though, because their input was anything but carefully curated. Steer well clear. - It seems to me this would not be an issue,technically, if AI training material was out of copyright in the first place. - @Jay The LLMs aren’t entirely useless, they just have very limited use cases. Some of those happen to be spectacular. Such as [https://en.wikipedia.org/wiki/Jacobian_conjecture](https://en.wikipedia.org/wiki/Jacobian_conjecture) and many similar things.Behind all this is a marketing ploy to justify a completely unsustainable level of investment in the area, as the aging tech firms look for another “growth’ story to pump up their valuations. None of that is intrinsic to the technology, it’s just a grift that’s grown up around it. Ed Zitron is good on that side of the story. [https://www.wheresyoured.at/the-subprime-data-center-crisis/](https://www.wheresyoured.at/the-subprime-data-center-crisis/) - In my opinion copyright is the worst reason not use AI for writing. If this legal interpretation survives it is bound to end copyright for everybody. Arguably every written work is an amalgamation of the works that came before. Show me that ladle that only scoops from itself, Goethe said. Why else would reading be such an important part of becoming a writer, if not to let your own writing be shaped by good examples. The main differences seems to be that any human alive has read less than the language model, and thus it is us who should be having the copyright problem not the AI. More importantly there are better reasons not to use AI. To write is enjoy the awesomeness of the writing flow, to spend time with your characters, to reflect and think on paper and to grow as a person. Why would you let a machine take that from you? Humans write because we have a need to share the truth, Anne Lammott wrote in Bird by Bird. Even in a distant future when machines may surpass our ability to write we will still feel the need to share our very human feelings and and the only way to satisfy this need is to write. - I mentioned on Tumblr that current AI should be used like aconite or cesium: useful only in very small amounts under tightly controlled circumstances, because in bigger amounts it’s highly dangerous and toxic. Most older AI fits this bill, but current LLMs are more like a major delusion in gaming: “Ice cream is great for computers, spoon it right in!” - I would say the best way to combat this issue would be to use a old fashion type writer on paper now tell me I used AI. - I would say the best way to combat this issue would be to use a old fashion type writer on paper. - I one hundred percent agree with you. I DESPISE AI. I’ve written more than 50+ works — the bulk of which date waaaay before AI — and I love my –em — dashes; always have, always will. But all true writers have a lot of shit-work. Stuff they started, then put away, that they will never show a single soul because it just didn’t jibe. I find it interesting that these billionaires didn’t ask writers to send their old stuff for a nominal fee. Or they didn’t train AI off THEIR own skills — their little secrets that they could share with the common man — such as how to make money off the stock market, secrets of networking with the 1%; how to write a Venture Capital proposal or a successful business plan — you know, the sort of follow-the-dot insider knowledge that the 1% know, that we don’t; that has no real creativity to it. NO, they simply ripped off the raw creativity of people paid piddling for their work ( Trad Publishers make their fortune FIRST) and then sell it back to us, while destroying the planet for their Data Centers. Linda Carroll, a writer on Substack said that having AI write for her is like asking it to “eat Ice-Cream for her” and I’ve felt that no one else has put so aptly what I feel about AI. Writing creatively is cathartic for me. A fundamentally impatient person, I got so mad at the submission/rejection/submission/rewrite/rewrite/rewrite process involved in traditional publishing that I jumped on Indi-publishing via Kindle as soon as it came out in 2010 and I’ve been doing it ever since. I made some really bad covers to begin because I stopped using the hated Blue Screen Of Death people…A writer’s bane. Remember them? I’ve used Google Docs from the beginning and had to train myself OUT of the habit of automatically saving every few sentences — which one had to do with Microsoft in the early days. So one of the instructors at a Digital Certification course I took laughed at my covers. But I felt proud of my books so she can go…impregnate herself. I kept at the learning curve because I could never afford to hire people and now my covers are fantastic. But these greedy tech billionaires are diabolical. The law is that whatever you write that is original is automatically copyrighted. You can register it, for a fee. But you don’t need to, in order to protect yourself. UNLESS…. You’re a stupid judge who doesn’t understand that AI was trained on millions of Indie-published works (not registered) and too, some Traditional publishers DID NOT register authors’ copyright. Theft is theft. The theft still took place. But those of us who built a catalog of Indie-published works — a billion-dollar industry, mind you — are exempt from receiving compensation. Does that seem fair to you? I was always gifted at writing and drawing. But I trained the hard way — under strict teachers and a professor who would say: “A Writer Writes!” And never give 100% for an essay, no matter how polished, citing that no writing is perfect. Yet right now, some Jabroni can have a machine spit out what was originated by me, and with enough money to create an initial burst of sales and reviews, gaming the algorithm enough to have a “best-seller”… It’s happening now! While the true writers have to work hours making an impact organically. Believe you me, if I had the funds to pay $1500 (quoted) for my work to be a ‘guaranteed best seller’ I would do it! There is AI now trained on the common themes of Romance genre and feeding it the variables (names of characters, etc) will regurgitate a romance for you with an eye-catching cover (trained on what sells most) to feed romance lovers. It’s “familiar”, yet different enough to sell. Not complaining, because AI written prose makes me sick. I get fatigued reading even emails from “gurus” who’ve decided to go that route and blocked every single one of them. I can tell a raw human voice from the superficial polish of AI. If you can’t be bothered with the effort of writing why should I read it? So I do feel that human writing is going to become more valued, more expensive and more rewarded, in time…. - @ Thilo Gross: “Arguably every written work is an amalgamation of the works that came before. Show me that ladle that only scoops from itself, Goethe said.” That’s not a terribly solid argument. Sure, our writing is all shaped by what we’ve read. But it’s refracted through the prism of our own experiences, thoughts, beliefs, periodic mood fluctuations, etc. It’s not a simple probabilistic regurgitation of a mishmash of the things we’ve read (at least when it comes to decent writers). “The main differences seems to be that any human alive has read less than the language model, and thus it is us who should be having the copyright problem not the AI.” The main difference is that human learning and information assimilation is not the same as a machine digesting enormous electronic training sets, then replicating patterns based on mathematical functions. I’m far from an expert on the topic, but you can look up pretty good summaries of it online. The copyright question, to me, is not really a question at all. “AIs” are not writing fiction, at least not yet. The situation covered in OGH’s post is that of a human being prompting an LLM to create text based on a pool of copyrighted text by other humans. Neither the human nor the LLM in this case have “created” anything, and you typically can’t copyright something you haven’t created. What would be interesting to analyze is a situation in which an LLM autonomously decides to write fiction. But I suspect that the outcome would be the same. - I fucking hated Clippy. In fact, he was the reason I first acquired a foam rubber brick specifically for flinging at the computer whenever it tried to help me. There are plenty of FOSS tools that don’t have builtin AI. Everybody first thinks of LibreOffice, but I find it slow and arrogant. I once had trouble viewing Word-created files with anything other than text in them. I sought help from The Document Foundation and was told that I needed to change the way I worked so that it conformed with the “LibreOffice way” which basically means never type anything without giving it a style. Do not ever just highlight something a choose bold. That’s not the One Right Way. The docx compatibility is godawful and they don’t seem interested in improving it. They blame Microsoft and Microsoft does deserve some blame for not doing proper documentation of that file format, but every other free office suite I’ve tried is a LOT more compatible than LibreOffice. Because, you know, that’s not the One Right File Format. That would be .odf. It is also ponderously slow to start. I’ve been using FreeOffice recently and like it rather a lot. - Ironically, I read this while waiting for Originality to analyze a piece of writing I need to turn in, only for it to tell me I’m a toaster. (I’m not a toaster, and I did not use AI to write that piece.) AI “checkers” really are as useless as the AI they’re supposed to detect! And before anyone asks…I’m using Originality because the client requires it, and I can’t just tell the client to fuck off because I desperate need the money. The past year has been ROUGH for anyone trying to make a living as a writer. Sadly, this $200/week gig is the best-paying job I’ve had all year. - [Deleted because hello, probably, Claude – JS] - I use AI. For figuring out orbital mechanics. Or doing “what if” research on a mechanical/structural physics question with regards to a space station or a ship under high acceleration, or something that’s beyond my innate ability to figure these things out on my own. I’m old enough to have a slide rule, but never learned how to use one, so… AI fills that gap. But WRITE MY STORY? HEEEEEECK no. Why would I give that fun away? It’s not about producing work, it’s about writing down my thoughts (which, if happen to be storyific, 1000x better). It’s frickin’ THERAPY and I ain’t giving that up, no way. As also a programmer … dang, if y’ain’t seen what AI has stolen from other places, y’ain’t been paying attention. - So, AI is yet another way in which young people can feel the odds are stacked against them. - Welcome to the day when a “100% Organic” label on a book has nothing to do with forestry. - Ah… and here we go. Already a first-time commenter with Impeccable Credentials That Can’t Be Verified has responded to an anti-AI post with a chunk of obviously AI-bot-generated drivel. It’s pretty impressive that it took 30+ comments to get there. - Human editor says: assume “and don’t think I *haven’t*already suggested so to my film/TV reps” because that makes much more sense? But the blip is 100% a human blip! - 100% agree with you. What is blowing my mind at the moment is folks are using AI to write “Fanfic” and posting it to fandom sites like Ao3, or so I’ve been told. Is that even legal = I mean there is a copyright loophole for fanfic under the fair use clause, but is it inclusive of AI? I don’t think so. And why would you want to use AI to write “Fanfic”? Doesn’t that kind of kill the whole point of doing it? There’s also a guy in the Ukraine who is using AI to translate his posts into English – he tells you up front that he’s writing the post with help from AI translation. (They are well written and somewhat poetic posts – which is…disturbing to say the least. But I do catch the errors in the AI in his responses to comments. There’s a limited number of ways AI can correspond with people, so it becomes obvious pretty quickly.) All of this is bewildering to me. I don’t like AI (I’ve seen all the Terminator movies and the series, etc – have people not watched or read science fiction? AI doesn’t come out well in it.). And find it annoying. But my work place has embraced it, and they are now writing contracts and legal documentation with it. The number of high-paying middle income jobs it will eventually do away with is staggering. - Zyvo Edge University [https://7d85c9sc1w8n0z9mct0bf5lo7g.hop.clickbank.net](https://7d85c9sc1w8n0z9mct0bf5lo7g.hop.clickbank.net) - John, you said you are comfortable with “trusting that readers understand it’s referring to Generative AI.” Umm…. Your base readership undoubtedly do understand this. But many people don’t, and all kinds of people see your stuff in re-posts and search results. Please, please say “generative AI” in the title or first sentence of your posts. It’s a tiny effort that could have a big effect, especially if multiplied by all communicators. You’re one of the better ones, because you include writing context right up front. Maybe you can explicitly encourage writer friends. In all cases of ambiguity (and AI is a doozy!), anyone with a platform can help educate people. For instance, my preferred news feed topics include both electronics and pro-democracy politics. If I see “resistance” in a headline, I want the context to be made apparent by the other words. Related communication issue, but off-topic: I’m amazed at locally-based news and community sites with page headings that don’t identify their location. Sites have gotten traffic from search for decades now (and not just to your home page, folks!) I live in one of the Portlands, and too many searches (even when I include “Oregon”) give results about the other Portland. I have to poke around and guess, look for proper names — not due to bad search algorithms but bad information on the sites. - The problem, of course, is how are you going to prove it’s not AI written or assisted. I’m a lawyer by profession with a penchant for writing. Sometimes I’ve been told that something I wrote as a hobby has been AI assisted. I tell them it isn’t but you cannot prove a negative. This article highlights dangerous territory. I can copy somebody else’s original work, claim it has been written by AI, they sue me, I give the AI defence and somehow somebody has to prove their position. Usually whoever sues must prove their position – but should this be a case where the ones of proof shifts to the defendant? But how can anybody prove something has been written or assisted by AI? It’s a legal wrangle not even Claude himself can solve. - I must admit that I use AI when writing. But not for giving me ideas or write, even small parts. It’s entirely mine. But I send my text et discuss consistency, style, or whatever. As I am not published, so far, I never realized it could be an issue. Considering your article, I may change my way of working. Thx - AI accusations are the new bocche di leone. They also serve as an IQ test — people who are radically anti-AI fail.