[Editing here to note that the HWA has removed this board member as of this afternoon. I’ll post the full text of that in the comments below.]
Today I, and every other HWA member, received the HWA newsletter in our inboxes. In that newsletter was an essay (“essay”) of pro-AI propaganda that made a number of specious points in an effort to, I guess, get us all to stop judging those who might use AI in their writing work and careers.
This is, of course, incredibly disappointing to see. This is a professional writer’s organization. It is the Horror Writer’s Association, not the Horror Prompter’s Association. A writer’s organization should be geared toward the act, the art, the craft, and the business of writing, not slop-wrangling and prompt-engineering using the shitty sparkling autocorrect of LLMs. It is grotesque to have pro-AI (“AI”) propaganda sent to us, the people whose necks are first under the boots of the AI companies. We are the people that have had all our work stolen immorally and illegally. We’re the ones turned into Soylent Green, so it’s a little fucking rude to have an essay piped hot and fresh into our inboxes that dares to be mad at us for criticizing anyone who eats Soylent Green. We’re getting cannibalized, turned into sausages, and you’re mad that we’re mad about it.
And when I say “you’re” mad, I don’t necessarily assume this is the Official Word of the HWA — already I’m seeing the rumbles that it most certainly is not, that they’re looking for ways to correct this. But the damage is done. The author, Sèphera Girón, is a trustee of the HWA, and saw fit to write what reads ultimately like an ill-thought-out tantrum and not a well-reasoned (if still disagreeable) pro-AI message. And so, regardless of how official this is, or isn’t, it’s disappointing that it ever made it out the door in the first place.
The “essay” is not something posted online, but you can find it online. One spot is here, on Bluesky, posted by author Holly Lyn Walrath.
On the one hand I don’t even know if the essay is worth legitimizing by taking it to task. On the other hand, it seems that its POV isn’t even “wow, AI is so great,” and more “wow, stop being mean about AI.” It attempts to take to task so-called “gatekeepers” and “bullies.”
So, I’m just going to go ahead and take some of the pieces from this essay and talk about how I feel these things are, at best, misguided, and at worst, poisonous notions about how stories are told and art is made.
“Technology keeps changing, oh-so-rapidly these days. We went from screaming about having to pump our own gas to screaming about self-checkouts. Back in the day, film studios were upset at the invention of home television sets, musicians were upset about electric guitars and synthesizers, filmmakers were upset about YouTube, visual artists were upset about photography, photographers were upset about Photoshop, and now writers are screaming about AI.”
I don’t know that pumping gas and self check-outs are relevant technological innovations here, but okay. But talking about film vs TV, electric instruments, photography, and so forth, leads into an argument that pro-AI (“AI” — again, we’re talking LLMs and what Tobias Buckell calls LIMs, Large Image Models) really like to use a lot, which is, “You stupid Luddite, you’re just afraid of the new technology.” And then sometimes they say something about horse-and-buggy versus the car, or some other false equivalency.
(Also, please look up the actual story of the Luddites.)
The reason all of this is false equivalency is that prompting your way to art or to writing is not the same as having a new tool that allows you to make art or tell a story. AI, here, is not a tool —
It’s not allowing you to do the work. You’re allowing it to do the work. Which is to say, you’re the tool.
The AI makes decisions. (Not literally. AI isn’t smart. It knows nothing. But programmatically, it takes pathways based on the chaos of its code.) You don’t make the decisions. You may guide them. But you don’t make them.
Look at it this way — if I have a camera, and I want to take a photo of a butterfly, I can do that. I can grab my camera. I can find a butterfly. I can track it, trying not to startle it, I can choose how to frame the shot in the chaos of that moment, I can choose how close I’m zoomed in, I can take one photo or a hundred. Then I can take the photos that I chose to take and do more work — cropping, adding contrast, modifying sharpness, whatever. Or I can take the photo and then choose to paint it, instead, using it as reference. Whatever. I have a lot of choices there, and they are choices born of the effort — and that multiply with greater effort — that I made out in the field.
But AI isn’t any of that.
AI is, metaphorically, me handing the camera to a robot and saying, “Go get me a photo of a butterfly.”
And then it brings one back. Maybe it’s what I wanted. Maybe it isn’t. But I didn’t make any choices except to fetch a digital errand boy to go make art for me.
“Meanwhile, the invisible oligarchs who control everyone are having a field day. Everyone is falling into their traps.
The real enemy isn’t your buddy who is using AI to make a book cover; the enemy is the Powers-That-Be who thrive on dividing everyone, especially in the United States. Look at how divided everyone in the US got over politics over the past ten years, and it’s now trickled to other countries. No more, are you a fellow human? Nope, it’s who did you vote for? Well, then you must be a hateful asshole. Instant block.“
I recognize here a snippet of truth that we should be aiming our cannons at the giant companies and oligarchs who are fucking everything up and not individuals — that’s entirely fair. At the same time, that doesn’t mean going to bat for the oligarchs. Sure, there’s no ethical consumption under capitalism — but there’s also ways to not gleefully be unethical. There’s “I reluctantly have to drive to work every day, which uses fuel and harms the environment” versus, “I get my Ford F-One-Fuck-You and I roll coal and drive aggressively and vomit fumes everywhere because I don’t give a shit, you woke beta.”
Thing is, even that metaphor kinda sucks — because in most parts of this country, driving a gas vehicle is still often a grim necessity. But using AI to write or make art? Is not. It just isn’t. You can write stories and make art without it. You could eschew its use last month, last year, ten years ago, a hundred, a thousand. It’s absolutely non-essential. (I’ve never used it.)
And since she wants to bring up politics —
I mean, yeah, if you voted for the hateful asshole — and have not yet had a Come To Jesus moment about that vote and that support — and I’m gonna wonder if maybe you’re a hateful asshole, too. It’s like, if someone kicks a puppy and you clap, no, I don’t automatically assume you’re a puppy-kicker, but I sure don’t think you’re someone I’ll hire as a pet-sitter, you feel me?
“The past few weeks have been so tedious with all the gatekeepers out in full force. They are whipping people up into a frenzy, and more division is happening when it comes to AI. Of course, the irony is that the gatekeepers are denouncing AI, yet they are doing it on Facebook, which uses AI. They are likely typing their pretentious screeds on their phones, which also use AI. They likely sell their books on Amazon, which also uses AI.”
Again, a deeply strained logic here. Choosing to use AI in your work is a choice different than simply existing in a world where it is being used around you and by some of the services. And honestly, most of the people who are on Facebook or who have books on Amazon are still doing their best to never use AI. They turn it off when they are able to do so. They avoid its use. They hate when they see it.
“Oh, but Sèphera, there are different levels of AI use, and some are OKAY, but some ARE NOT.”
I mean, yeah? Isn’t that life? Everything is on some kind of spectrum, and sometimes that spectra is, in fact, in an increasing gradient of “enh” to “demonically terrible.” We’re all making choices out here however we can. AI that finds cancer or gets me to the mall faster is different than the one that stole my work and shits it back out when prompted.
“Are you watching TV? Are you using Alexa? Do you have home security? What kind of car do you drive? Do you go to the doctor? Do you watch TV? Do you go to the movies? Do you have shopping apps on your phone?”
What?
What does this even mean?
Am I participating in society, you mean? I’m not new here. Yes, we’re all doing these things all the time. This is very shallow and quite silly. “DO YOU BREATHE AIR? DO YOU USE THE TOILET? ARE YOU CLAD IN FLESH AND ACKNOWLEDGE CORPOREAL REALITY? AH! THEN YOU EXIST IN THE SAME PLANE OF EXISTENCE AS AI! HYPOCRITE!”
Like, what are we even doing here.
“Oh, Sèphera, that’s not the point. The evil is GENERATIVE AI, not all the other AI!
Sure, sure. Let’s somehow separate one form from another.
Pick and choose your evil so that you can keep gatekeeping and let the oligarchs win as the little people continue to fight against each other.”
She’s very close here to getting it, actually.
Yes, you can separate one form from another. In fact, you really should, because that’s how reality works — different things are in fact different. Worse things are worse than better things!
It’s like saying that eating food includes eating a Thanksgiving turkey and eating your neighbor, Dave. Well, they’re both eating! Oh, what! We’re going to separate one form from another now!! Way to gatekeep my cannibalism, you bully!!!
“Yes, we know AI is using up resources. So does driving a car or a plane. So does building a new home. So do a lot of things. Humans have always used up Earth’s resources and usually not nicely. There are reservations and other Indigenous communities in Canada that have never had clean water and have been ravaged by pipeline construction. But somehow that’s been okay for the past fifty years or a hundred years, and very few people scream about that. But suddenly, AI is the enemy, not everything else we’ve been doing all along.
Suddenly YOUR water might not be so clean, so now it’s a Big Deal.”
“Humans have always used resources, and indigenous people have always been fucked over, so oh well” is not neither a nice nor particularly realistic argument. “Nothing can ever be better, so make things worse” is essentially the thrust of it.
“All the drama and arguments in the horror writing field would be less prevalent if people could have conversations and educate each other. No one really and truly knows what’s going on or even what the bottom line or end goal is. But what we do know is that here at the HWA, we are all writers, we all work in publishing in some capacity, and AI is here to stay. HWA has a public stance against generative AI. But not everyone knows what that is or even how to use it.
Instead of blocking people and then telling others to block those people because those people DARED to explore AI potential for their own careers, why not be human and educate each other? Watch out for each other. Help each other. RAISE EACH OTHER UP.“
I don’t really even understand what all of this means, and honestly, re-reading her screed just makes me both annoyed and tired.
Here, I think, is the thrust of it.
Her essay behooves us writers to be “human” and to “raise each other up.”
Except, AI is not human.
It’s made of humans! It took so much of what makes us very interesting and chucked it all inside a fucking Instant Pot and let it simmer until it’s all just steamy goo. So, seeing people use it — meaning, scooping up ladlefuls of material stolen from their peers — sure doesn’t feel like them raising us up (except to their greedy lazy pig mouths), so why are we obliged to raise them up? It’s the “tolerate my intolerance” with an AI-friendly spin on it.
None of this is new from me, but I think it really behooves people to understand why some folks are so very angry at AI, at its creators, and at those quislings who dare to use it.
It’s very serious to us — a theoretically-existential threat! — to have our work taken without our permission and used to feed gluttonous software designed by techbros who literally say the quiet parts out loud when they talk about how it’ll destroy jobs, how we’re all gonna have to move to the fucking moon or instead live in a bunker beneath a data center somewhere, how if they actually had to pay creators for licensing rights they’d be bankrupt. They steal our work, they throw it into the machine like so much coal, and then use it to enrich themselves while simultaneously harming the environment and poisoning the fidelity of all our information systems. It is very clear that they want to treat information and knowledge as a resource to be mined — a utility to control, one that can be leveraged to extract payment from us. We and our work is all just material — Capital-C Content — to be melted down into slag.
So yeah! We’re kinda fucking mad about it! And when we see someone we like using it, we kinda want to kick some sand onto your shoes. You look like a sell-out. Like someone who just sold his buddy to the Cannibal Farm down the road. When someone like Hank Green is found to be using AI, it’s not that I think Hank Green — who by all reports seems like a lovely guy! — should be strung up by his heels and hit with crowbars. But it’s still disappointing. Because he’s a science communicator whose whole schtick is to answer questions and provide some measure of entertaining truth while also being someone interested in and invested in the health and wellness of our world and society — so when he’s seen using technology that sure seems to be damaging the health and wellness of our world and society, meaning it’s harming the many to benefit the few, well, yeah, you’re going to see people wanting to take him to task for that.
All this goes back to the fact that the HWA — and the SFWA, and the Author’s Guild, and so on and so forth — are professional writing organizations.
And using generative AI is not only not writing, it should not be a part of writing, and it should be viewed as antithetical to our crafts and our careers. And before you say “but what about autocorrect,” know that you and I both know what I’m talking about when I say generative AI or LLMs and it doesn’t mean using fucking spellcheck, okay? Okay.
If you wanna use AI, I can’t stop you.
But I can definitely give you a look like you just giddily shit your pants on a crowded train. That’s not me bullying you. That’s me acknowledging you shat yourself willfully and with joy and now we’re all having to deal with it.
You make your choices.
As for the HWA, I sure hope they’re going to walk this back and try to ensure something like this does not happen again, because it’s hard to stay a member of the organization if they can’t keep something so simple as their newsletter under control. Anyway. Have a great day. And fuck AI, amen.
conniejjasperson says:
Well said, Sir.
August 5, 2026 — 11:52 AM
BradentonDeb2021 says:
Well said, indeed. Sounds like “Kilby Blades,” the person who single-handedly killed the original Nanowrimo, is back. With their faulty reasoning in favor of AI, they are ruining it for all of us.
August 5, 2026 — 11:56 AM
Melanie Cossey says:
It so reminded me of that too!
August 6, 2026 — 2:40 AM
kathleea says:
I recently joined HWA and to see this in my inbox makes me wonder if I did the right thing by joining because I am vehemently against AI. It steals our words and wants praise for doing it. No. Just no.
August 5, 2026 — 12:01 PM
A Writer Who's Had It says:
Yes. Well put. But let’s not forget an awful lot of white authors are getting away with using LLMs while authors of color aren’t. Look at Charles de Lint. Two days of some pushback, but none of his close colleagues spoke up. He’s fine.
That’s what the writer of the HWA screed is hoping for. Tacit acceptance for bad behavior.
August 5, 2026 — 12:04 PM
Jeff Dwyer says:
I’m with you on AI. It’s a dangerous weapon in the wrong hands. Anthropic for example. One of the biggest, most intrusive and secretive AI-panderers, it has been found that and has admitted that their goal is to scan in, assimilate, and then DESTROY ALL printed books, including historical artifacts and irreplaceable documents. This was in a verified internal memo to staff, supposedly secure against disclosure.
They want to own all human knowledge, and take the right to tailor it, disclose or hide it, use it, and sell it for ANY legal or illegal or unethical purpose the buyer has for it. The U.S. Department of Defense has demand unrestricted use of the Claude AI model. So far Anthropic has refused an ultimatum to drop safety guardrails against fully autonomous lethal weapons and mass domestic surveillance. We’ll see how much $$$$ and threats from the MAGANAZI regime it takes to eliminate their refusal.
In the arts, AI is catastrophic, in many ways. Musicians, writers, and painters are already up against the wall trying to survive. This technology steals ALL their ideas and craft, mashes it together, and spews out reasonable facsimiles, with only a simple prompt by the fraudster ordering the product. A pig farmer could now write War and Peace while out slopping his hogs, and sell it for millions. Same with music and art. As a writer and a musician, when I encounter this product, I am impressed by its command of language and technique and its slickness, but I am stunned by its total lack of humanity, insight, or creativity.
And lets not even get into the unethical, fraudulent, libellous political uses of AI we are already being barraged with.
I’m not a fan.
August 5, 2026 — 12:04 PM
Will Humphreys says:
Cory Doctorow recently wrote a great article that might cheer you in The Nerve online newspaper. Highly recommend. Can’t copy a shortened link for some reason but should be easy to find.
August 5, 2026 — 12:19 PM
jenesuispasdave says:
https://www.thenerve.news/p/cory-doctorow-ai-business-incentive-to-lie-useless-chatbots-report-nikhil-suresh
August 5, 2026 — 7:22 PM
Mike Cook says:
“And fuck AI, amen.” — works for me.
August 5, 2026 — 12:26 PM
Jennifer Loring says:
JFC. Just another reason I’m glad I left the HWA.
August 5, 2026 — 12:29 PM
Michael Patrick Hicks says:
Oh my god, same. Also, it’s great to see you here, Jennifer! Been too long; miss you on Bluesky!
August 5, 2026 — 2:49 PM
Hook says:
Beautimous Chuck. Hard to believe that newsletter was released. I had trouble reading Holly’s stuff, considering her position, but enjoyed your perspective, rants, and agreements, as such. Maybe HWA should allow you to write a counter newsletter, “The Demon’s Advocate” issue to discuss con-AI less-than-pro-paganda. As a writer and an artist/designer, I have double the issue, double the gripe, and it really doubles me over at times having to trudge through AI pro and con reads. For those who are about the craft, it’s a logical understanding. Those who are not, a flakey shortcut. But as AI grows in rogue-liness, I’m sure humans, in general, won’t matter for opinion.
“Come with me if you want to live, to think, and make your own decisions.”
– AntiTerminator
(Spell-checked by m’own brain)
August 5, 2026 — 12:29 PM
Susan Spann says:
Fuck AI and the horse it chewed up to get in here. World without end. Amen.
August 5, 2026 — 12:41 PM
Naysh says:
Thanks for writing this.
August 5, 2026 — 1:33 PM
Ali says:
What I want to know is how this piece (hers, not yours) met editorial standards without some kind of dense, pointed disclaimer. Because that is also a problem!
Really great post, Chuck. AI can suck a tailpipe.
August 5, 2026 — 1:48 PM
Faith S says:
As a songwriter, I want AI to do my laundry and dishes, and to clean my house, while I go off and write songs. I want Rosie from the Jetsons, not Data from Star Trek. By recent estimates, 30-45% of Spotify’s daily uploads are AI-generated songs.
August 5, 2026 — 1:50 PM
Arlan says:
Completely agree, Chuck. I’m an engineer. It would be nice to use AI to give me ideas for a better design. As you said that would be simply stealing another’s ideas. It would be nice to be able to trust AI to look up specifications, tolerences, test results, etc. However you never know when it’s going to lose its “mind”, spew out crap and suddenly the bridge falls down, or the plane falls from the sky for no ther reason than someone was too lazy to actually look something up from verified sources and do the actual work to design or test something.
I use computers all day to make my job faster and more efficient, but as you said, it’s my thinking that comes first. AI leading my work should never happen…EVER.
I’ll get off my soap box now…
August 5, 2026 — 2:15 PM
Michael M. Hughes says:
I think this is important for context: The author is either partnered with or otherwise making money off studio.com, an AI-powered writing coach app. So the anti-AI folks (like me) are harming her attempts to sell the app.
See: https://studio.com/apps/sephera/writer
August 5, 2026 — 2:15 PM
Bryan Schmidt says:
I think HWA needs to make a very clear official statement about this matter. Didn’t they just post an official AI policy? they have to have a clear statement because you can’t have any confusion moving forward.
August 5, 2026 — 2:23 PM
rpulfer says:
A fun read as always, Chuck.
As someone who’s on the fence on AI, I think this HWA essay does a really poor job of defending AI. Inadvertently defending “shitty publishers” while also defending artificial intelligence is…a choice?
And can we stop calling anyone who disagrees on AI a gatekeeper? None of us have that kind of power. My friend was shocked SHOCKED that I didn’t think he should use AI to write the novel he’s been wanting to write for years. I’m not a gatekeeper – I just don’t think talented writers should ask AI to write a novel in a single prompt and then put it on Amazon in the next breath.
But I do disagree to some extent, Chuck. I’d be lying if I didn’t use AI here and there. Mostly, I use it to have a conversation about my projects. It helps me brainstorm, helps me find connections in my writing. But while I use AI as an alpha reader at most, I keep it entirely separate from my writing and editing process beyond that.
And I do get the fear this author has about being shut out and de-platformed for having a different opinion on AI. That’s a very real and credible fear I share too.
I don’t know the extent to AI is harmful and beneficial. I don’t know if I’m doing the right thing, nor do I know how AI will shake out in the creative space one way or the other. All I can is make the best decision going forward.
I can’t stop you from looking at me like I giddily shit my pants on a crowded train.
All I can do is make my bed the best I know how and lie in it.
Now if you’ll excuse me I have to go change my underwear.
August 5, 2026 — 2:44 PM
Tanek says:
People are certainly allowed to have different opinions about “generative AI” and its use. But it has to be acknowledged that someone’s opinions can sometimes be used as an indicator for what kind of person they are and can help you decide what kind of personal or professional engagement you’d want to have with them.
If you can look at how the LLM databases were built, if you can listen to the company execs go on about how all this creative work was essentially “worthless”, so they should not have to pay for it, if you can hear teachers and professors and editors and publishers explain how the genAI use is affecting students and minds and work product, if you can read the reports about AI datacenters taking water and electricity away from communities, if you can understand all of that and still be on the fence as to whether it should be used for creative work today?
Then I guess all I can say is, I hope you have a decent supply of pants on hand.
August 5, 2026 — 5:51 PM
Richard Pulfer says:
Fair enough. Just to be clear, when I say I’m on the fence, I’m saying I remain “skeptically optimistic” about the promise of AI. I understand not everyone feels the same way, and I respect that.
I disagree with 90% of what AI execs say. I don’t think creative work is worthless, nor do I think that AI should be used to replace or exploit the labor of teachers, professors or freelancers. In fact, one of the biggest reasons people say “Fuck AI” so much is because of overreaching, condescending BS that has inevitably escaped the lips of many a corporate exec.
This might be a contradiction, but I don’t support data centers…at all. I hope AI gets more environmentally sustainable in the future, but until then, I accept one can use a cell phone and still not desire an electronics landfill in their forest preserve. And given the choice between AI and nature, I’ll choose nature every time.
All that said, I think AI can still be helpful in creative work. I use it mostly to start a conversation about my project. It helps me develop the themes of the story, realize connections between plot elements. I do NOT use AI to write my book. I write the story, I do all the editing, I create the characters, I do everything . AI is just the assistant – not the replacement creator.
August 6, 2026 — 10:30 AM
mystiberry says:
Perfect. Thank you.
August 5, 2026 — 2:51 PM
Mur says:
I’m not a member and have only read replies, but what kind of “gate” are we talking about here. I’ve understood editors to be accused of gatekeeping by not letting newer writers in, or established writers don’t introduce or help newer writers, but what gate is kept, and who’s hurt by it? (Fills in “AI is here to stay” on the bingo card. I hope to hang it next to “NFTs are here to stay.”)
August 5, 2026 — 3:15 PM
ina says:
Her nonsense is just another version of confusing not being allowed to make choices and being criticized for the choices you make. It’s very much what the US right wing loves to do (e.g. first amendment stuff- “cancelling people in censorship! eek! eek! mah rahts!”). You use AI, you take the consequences, and if those consequences are that people don’t read your stuff or criticize you openly or encourage others to not contribute to your selfish damaging of the environment, well…gosh, welcome to preschool, where actions have consequences.
August 5, 2026 — 3:18 PM
Michelle says:
The author of that essay must have gotten bribed by AI marketing firms to write that. My mom is retired and in her 70s, had ample time to watch the nightly news. And I’ve had to point blank explain to her how sponsored content works multiple times now, that this isn’t the evening news of the 60s and 70s back when we had the Fairness Doctrine. Everything positive she thinks she hears about AI is a lie, a hoax and a paid advertisement. And yet she says to me, “If you want to stay in your career, you’re going to have bite the bullet and learn AI.” Like it’s a Word update I can’t contend with because I’m stubborn.
I used to write articles, it’s all been STOLEN. There aren’t enough jobs to go around anymore to use AI skills in. An old editor of mine who knows all the AI skills still got laid off earlier this summer. I haven’t signed a client in 3 years this month and not for a lack of trying. I had to move over to a stressful healthcare job while I post my writing on the side. At least it’s not AI.
August 5, 2026 — 3:37 PM
writerSeth says:
I’m feeling extra spicy about AI since I accidently bought an AI prompted audio-book over the weekend. Stupid thing can’t keep track if its in first or third person between sentences in the same paragraph. Refers to major plot points as if they’d been mentioned before but now is the first time it’s bringing it up. Weird asides, shifts in character voice that make no sense, odd repetitiveness. Utter garbage.
I returned it and feel zero remorse. Only downside is that I couldn’t leave a review after.
August 5, 2026 — 3:38 PM
Charlie says:
Another very important detail is that the author of the newsletter also sells and AI-based writing coach app. That seems relevant.
August 5, 2026 — 3:44 PM
Fatman says:
Most of the pro-AI screeds I’ve read online were written by people peddling some sort of AI “writing” app. People, I should add, without significant writing or publishing credentials.
Before AI, those same folks were scamming wannabe writers into pointless writing and marketing courses. It’s the same tired old song, just in a different medium. I’m just surprised the HWA allowed itself to get dragged into this one.
August 6, 2026 — 8:12 AM
Maria says:
The butterfly photo analogy is excellent.
I’ve been a professional writer (someone who actually gets paid to write) since 1990. I have never used AI in my work and never will. Not for research, not for ideas, not for anything. if a writer can’t be bothered to do all the work involved in writing something, why should anyone be bothered to read it? I respect my readers too much to give them anything less than what I can create on my own.
August 5, 2026 — 3:52 PM
gregh121 says:
Ironically, I blocked HER everywhere because of this mean-spirited bitchy ad for her app.
August 5, 2026 — 4:55 PM
terribleminds says:
Statement from the HWA:
Statement from the HWA Board of Trustees:
The Executive Leadership and Trustees on the Board of the Horror Writers Association has voted to remove Sephera Giron from her position as Trustee effective immediately.
The Horror Writers Association is committed to the promotion of Horror and Dark Fantasy literature and the broadening of intellectual and cultural horizons by way of good works in these genres. In support of this mission, the Board established a robust artificial intelligence policy, following input from all levels of membership. This policy emphasizes the organization’s sincere commitment to not only uplifting creators but also protecting a culture of labor compensation in the arts, aligning with principles that safeguard our community.
To this end, the HWA Board firmly and resoundingly rejects the opinion piece of Trustee Sephera Giron as included in the August 2026 installment of the HWA Member Newsletter in the “In the Trenches” column. This piece does not meet the standards of care and support to which we hold ourselves as elected members of this writer’s organization. We regret that it was published in our newsletter without proper vetting. The HWA will review all newsletter submission procedures, including those submitted by Trustees.
Further, we reject the assertions in the piece, particularly those calling authors with genuine and sincere concerns regarding the proliferation of genAI (more commonly referred to as simply AI) Luddites and/or gatekeepers. Additionally, in keeping with the organization’s ongoing commitment to the uplift and promotion of diverse voices in Horror literature, we firmly denounce the comparison of concerns over AI data center proliferation to the ongoing struggles over Indigenous water rights and ways. As the leadership of the Horror Writers Association, we sincerely apologize for the hurt that these words have caused.
We have heard your concerns and write today with a full understanding of the severity of the error made by way of this opinion piece’s publication. It is not in keeping with opinion of the HWA, the Board, and the general consensus of membership. In the future, we will do more to better to review and vet material included in the newsletter so that a repeat of this incident does not occur.
Given that the column in question fails to uphold the principles of the organization, does not reflect the consensus of leadership and membership alike on the subject, and demonstrates a failure to fulfill the responsibilities of a Trustee, the Board has voted to remove Giron from her position effective immediately.
We have always been an organization meant to protect writers and that is what we will continue to stand for.
—
The Board of Trustees of the Horror Writers Association
August 5, 2026 — 4:58 PM
Sarah Roark says:
Speaking as someone with two works in LibGen, I’d have to look at what this person wrote. If it was really a tantrum, then obviously that’s not okay. But you’re expressing a lot of misconceptions about AI in here and I’m sure a lot of pro-AI people would call it a tantrum (I wouldn’t, but I hold positions on this subject that neither of the spectrum’s extreme ends are happy with.) For instance, because of the way the tech works, which I recommend studying, LLM training data is totally atomized during the learning process. It’s “read,” not “stolen.” That’s why US courts rule training fair use. LLMs like 99.9999% can’t produce literal plagiarism, and are not going to reproduce anyone’s copyrighted characters, setting or specific plots by accident. They can only ‘fanfic’ on well-attested properties if *you* specifically tell them to, and even then the odds of them getting it too wrong to be much good out of the box are high. (Odds are better for successfully generating a really well-attested character in an image model, but again, you have to specifically ask for it.)
That doesn’t make downloading pirated works (i.e., that were behind paywall or had DRM broken) legal, so LibGen should be sued, and AI companies should be sued for obtaining such datasets. But aside from that, I think free publicly-available material should be fair game for training.
August 5, 2026 — 6:44 PM
Jeff Dwyer says:
Read up on Antropic and tell me you don’t think they are ‘stealing.’ Not only stealing, but systematically and intentionally destroying their sources.
August 5, 2026 — 7:24 PM
Sarah Roark says:
I have read up on them, extensively. Again, it’s been ruled in court that training AI is by itself fair use. But material can be obtained illegally. They WERE rightly sued and had to pay out a large settlement to authors and companies (unfortunately not to me; my books were all work-for-hire so I get screwed) because they used a well-known and at the time often-used data set (LibGen) that includes two pirate archives which contained many pirated books. I.e., stolen from behind paywall or DRM-cracked, thus not legally publicly available. That’s how they got my two books. That was wrong. But that was an offense of downloading a single copy of pirated content (or two; or maybe more, depending on how many of the archives compiled in LibGen had pirated content). They effectively ‘BitTorrented’ stuff for private use, like people do when they create pirated music libraries to listen to. Which is illegal. The use (training) to which that material was put, though, is ruled fair use so far.
As for the use of physical books, another court ruled that it was okay for a company to scan a physical book for their company data *so long as* they destroyed the original they had bought, because that didn’t add a usable copy to the world. I really don’t like the destruction requirement, because I’d rather the physical books were preserved. I think it should be legal to retain digital copies of a legally owned work for private use. However, there’s been unsubstantiated news coverage claiming they’re destroying *rare* books. First off I highly doubt this, because rare books are extremely expensive. Second, I haven’t been able to find a source for the claim. The most I’ve found was a bookseller quoted in the original scoop: he himself sells books to the companies doing the scanning, but he also worries that rare books are *maybe* being sold by somebody. Presumably not himself, because of course an honest interviewee would cop to that right? :-P…but y’know, *other*, less moral sellers. He did not allege it was actually happening, however. It looks to me like this quote was exaggerated into a fact.
August 6, 2026 — 5:41 PM
Debi Gliori says:
And how can anyone be chilled about Anthropic’s assistance to the IDF in the ongoing genocide in Gaza. Our work may, as you say, be ‘atomised’ in the training of a LLM , but if that same LLM goes on to be used as part of the machinery of war, then our atomized works are contributing to the murder of innocent civilians.
And I want no part of that. Fuck AI. Fuck Anthropic and their – oops, we got found out, we’d better bribe y’all into silence.
Not in my name.
August 6, 2026 — 4:00 AM
Sarah Roark says:
I’m pretty read up on Anthropic and I haven’t seen any allegations that their software is being used for the Gaza genocide. At all. That’s solely an IDF operation and they have their own model, “Lavender,” which they use for their targeting and such. What’s your source?
The *US* military is using Claude for I think intelligence and analysis?…but they’re doing that *despite* a supposed government-wide ban on Anthropic as a “supply chain risk.” Whucb designation they got because they refuse to allow DoD to use their products for autonomous weapons or domestic mass surveillance. Supposedly the DoD is required to wind down their use of Anthropic soon in accordance with the ban, but we’ll see if they actually do it. They really really WANT to use Claude because they know Anthropic makes great models. ^^ But in any case, Anthropic has maintained red lines on the military use of their products that Google and OpenAI have not. (Google’s stated that Gemini *shouldn’t* be used for autonomous weapons or mass surveillance but that’s not binding, and OpenAI has language in its defense contract saying its stuff should not be used for those purposes, but it’s toothless because that doesn’t actually legally prohibit it. Both companies’ employees have protested, calling on their bosses to impose the same restrictions as Anthropic.)
August 6, 2026 — 6:02 PM
John Smith says:
You’ve fundamentally misunderstood what ‘atomization’ does here. The data may not be directly visible in the parameters somewhere, but the ability to reproduce it at will still exists. That’s the whole point, that real plagiarism is not just possible, but rather easy to do. There have been tests of various LLMs to see if they can be convinced to produce quotes from particular books, and many of them will easily reproduce full paragraphs found in their training data. The fact that they only do so ‘when asked’ doesn’t make it ‘not plagiarism’.
Also, 99.9999% ‘not literal plagiarism’ is an odd statistic, since LLMs are famously unable to hit 6 nines of accuracy on anything. I’m concerned that your understanding of LLMs comes from sellers of AI LLMs, and not from researchers looking at what they can actually do. There are examples of them providing significant chunks of books, word for word, when asked (and not connected to a search API).
August 6, 2026 — 10:45 AM
terribleminds says:
Bingo to this.
August 7, 2026 — 10:08 AM
Sarah Roark says:
BTW, I say all this as someone who does NOT let AI create any part of my public-facing work (though I tinker with it in private, mainly to see how well it can work in realtime RPG game sessions, and I’ve also used it for brainstorming and feedback). Why do I not? For the perhaps unfashionable reason that I just don’t WANT to. I’m a writer. I like to write. So I write.
August 6, 2026 — 6:06 PM
Susie Lindau says:
“It’s not allowing you to do the work. You’re allowing it to do the work. Which is to say, you’re the tool.” Perfect. I hope your quote goes viral, Chuck.
August 5, 2026 — 6:56 PM
Diane Corso says:
Another thing making me question my HWA membership, despite my never having met a single mean person at Stokercon. Thanks for writing this well-considered response. I never read my newsletters, but the retraction they sent out made me actually read it, lol.
August 5, 2026 — 7:23 PM
Derek Murphy says:
Solidly reasoned, all the way through. There might be some good uses, like spellcheck, though that replaces editors, and which part is the author’s responsibility, craft or art, is a sliding scale that shouldn’t necessarily have moral conditions (does every book have to be Art? Are writers only permitted to be artists, and not proficient technicians, ie unlike nearly every other kind of gainful employment, are authors judged on the purity of their process rather than the quality of the output?) being mad at ai use is probably the moral high ground but why must this be a moral issue, and how do we deal with the current accusations being cast at ‘real’ authors who didn’t use ai but kept their m-dashes; this seems to penalize a certain prevalent style of writing forcing us to become performatively human to avoid becoming mislabeled or having our books run through ai by skeptics without our consent anyway, can’t we hate ai and also be uncomfortable with the witch-hunts? And since real humans are expensive to hire, publishing remains a pay to play industry that requires significant investment even though ai has allowed some authors to do some things cheaper with better results (do we maintain that no book should be published without a team of experts and the required budget?) most of my books are a decade old and I pay for a lot of services for landing pages or KDP analytics or spellchecks or formatting software, and I can build all of that for myself now and save $100 a month, is coding or commerce the noncreative stuff that might be permissible and who decides that; which kind of work has unreplaceable human value and which kind of workers do not deserve protection?
August 5, 2026 — 9:59 PM
Sarah Roark says:
Good points. One of my biggest complaints about anti-AI discourse (which has *some* critiques I totally agree with) is that it’s become a puritanical moral panic. It’s very binary, and people who use AI often get epic mass public smearings. Which is especially unfortunate when it’s someone who’s *falsely accused,* which is now happening all the time especially with visual artists. 🙁
Even those who simply won’t join the crowd in throwing stones and fantasizing about guillotines are called immoral or stupid stooges for the evil billionaires who literally want to mass murder people and destroy the environment. The discourse is totally poisoned by this, as well as by the doomers and some sadly prominent but bonkers pro-AI commentators…and almost no discussion of ethics or reasonable regulation happens as a result.
August 6, 2026 — 6:14 PM
terribleminds says:
It’s not “puritanical moral panic.”
They built systems that:
a) are environmentally damaging at a time when we really, really cannot afford damage to the environment
b) are built off of stolen material and stolen labor, provably so
c) work very hard to move money out of labor and into the hands of the wealthy
d) are very useful tools if you want to introduce a great deal of noise into our informational ecosystem, which is to say, lots of misinformation and disinformation, but also just noise in general (see the absolute spike in fake AI-generated stories and books submitted to magazines, to online marketplaces, or even just to the web)
e) are created and shepherded by people who seem very very comfortable telling us how this will destroy jobs (and it has, just look at today’s job numbers) and we’re all just going to have to live in the data centers now, sorry, they’re going to go their White Lotus resorts on the moon, bye
And that’s just the start of it. So, please don’t call this a “moral panic.” That is incredibly dismissive to the real concerns of real people, and incredibly conciliatory to the AI techbros and the slop-farmers.
August 7, 2026 — 10:07 AM
Sarah Roark says:
I apologize in advance for length. These are complex issues and I’m autistic, so I can be a lot to handle sometimes. Wouldn’t blame you at all for asking an LLM to summarize it. ^^
The puritanical moralism and panic of it isn’t so much in the actual substance of the critiques. As I say elsewhere in thread, there are anti-AI critiques/concerns that I share and find deeply valid and urgent. The issue is in how those critiques are *prosecuted and acted on*. ATM, the scene 100% cannot distinguish between actual bad actors and innocuous (or if you like, less harmful) use. Supernatural evil is imputed not just to the AI engineers…most of whom that I’ve met are very thoughtful and ethical, and a number of whom are lefties!…but worse, to the machines themselves, which is utterly stupid. People watch too many goddamn movies.
The AI companies are also judged right now in a way no other tech companies or industries are. AI is accused of being *intended* to destroy the planet and depopulate the non-elite classes of humankind (the latter a claim also made during the Industrial Revolution, btw). I even saw a couple of weeks ago a conspiracy theory that the companies building data centers don’t *really* want to run data centers, (because OBVIOUSLY AI is just a scam and everybody’ll stop using it any day now), but just do it to get the *land*, which they’ll use for something else after the data centers inevitably die. X_X We are of course talking about companies that if they want land can just…*buy land.* But this is the level the crazy has reached now.
This — the witch-hunting, the imputation of Satanic deeds and motives and not just physical but *spiritual* contamination; the obsession with purity vs corruption; the mass pillorying and shaming of people, especially creators, who use AI (or, increasingly, are falsely accused of it, because why wait for facts); and the spreading of blatant, lurid misinformation, are the hallmarks of a moral panic. One with a very characteristically American flavor of Puritanism.
As to your other points:
August 7, 2026 — 9:41 PM
Sarah Roark says:
a) Yes, data centers are environmentally damaging, but 1) new tech is coming online now and in very near future that makes it much less so, 2) much of the problem is in the power generation, which puts it in company with most large industry, and 3) we don’t accuse factory livestock raising, large-scale agriculture (which guzzles epic water amounts, destroys habitat, and pollutes the soil with nitrates) or LNG power plants of ‘destroying the planet’…even though they’re much bigger offenders. Hell, golf courses use more water than AI (especially if you exclude the power generation). If you refuse to ask ChatGPT a single question because Mother Earth but keep eating burgers, you’re causing waaaaay more harm than if you did the reverse.
Personally, I think it’s urgent that we mandate *very strict* environmental standards on data centers (esp on BYO green power, low- or zero-water cooling, noise mitigation, on-site treatment for any blowdown water if local systems can’t handle it, no shortcutting of community and environmental review), but bans and long-term moratoria are dumb. Especially since we’ll continue to need bigger data centers not just for AI but for literally everything else that happens online — including phones, cloud services, email, streaming, etc etc.
b) Explained above why training is fair use (although *downloading* pirated material to use in training is a violation). If you want more details about the atomization and why it matters both practically and ethically/legally, lemme know.
c) Would it help if I said that my firm conviction is that late-stage capitalism simply won’t survive the post-scarcity/automation crisis that’s been going for 50 years if not 80 (if not 250), and that AI is only pouring even more accelerant on? And I expect it to collapse quite soon in societal terms?
d) Slop IS indeed a huge issue. It’s been an increasing issue since the Internet came around, but AI obviously allows much greater volume. The only thing to do about this, as with the Internet, is better curation mechanisms, which will probably in part require AI. Fortunately, models continue to get better and better, so AI generations even from free/cheap models should be less and less sloppish as time goes on. Already there’s epic improvement from, eg, launch-version ChatGPT.
August 7, 2026 — 9:42 PM
Fatman says:
“It’s very binary, and people who use AI often get epic mass public smearings.”
Really? I’ve only seen this happen with folks who publicly try to pass off AI “creation” as their own original work.
Personal use of AI is your own thing, and will depend on how you feel about AI. But it’s unlikely you’ll be targeted for a witch hunt simply for creating funny cat videos on your computer.
August 7, 2026 — 10:20 AM
Sarah Roark says:
I just saw a prominent YouTube creator (Hank Green) get absolutely shitpiled, not even for writing scripts or making graphics with AI, but just for using an LLM to do *initial* research (which is like using Wikipedia for such: very useful as long as you verify everything, and he knows that so I’m sure he did) and brainstorm. He had to do a humiliating public self-flagellation about it because that’s the blood people demand nowadays.
August 7, 2026 — 7:07 PM
Fatman says:
Apologies if I’m poorly informed about this – but didn’t Hank Green try to present AI-generated research as his own? Or at least initially fail to disclose AI use, and only did so after pressure from followers?
Which would very much fall under “publicly trying to pass off AI ‘creation’ as their own original work”?
August 7, 2026 — 8:24 PM
John Atherton says:
MAN, you should really not hold back! Come ON! Tell us how you REALLY feel about AI and shit. SO STIFLED!
August 5, 2026 — 10:33 PM
s3lkie@gmail.com says:
using ai means taking advantage of other creatives. it means supporting a system which kills creativity as well as the capacity to earn anything. it means fucking over other creatives.
fuck ai.
August 5, 2026 — 11:26 PM
Gordy T says:
Hey Chuck, I just got hit with a sales pitch from someone who thinks I should be using their new Dragon Romantasy GPT prompter. Their reasoning? Quote from the pitch: “The trouble with trying to write a novel from scratch is…
The process of writing a novel forces you to solve too many problems at once.
You are building the world.
Developing characters.
Managing the slow-burn romance.
Planting secrets and betrayals.
Planning reveals.
And somehow, while doing all of that, you are supposed to write compelling prose.”
That’s it, a writer has to solve too many problems in one shot and write something worthy of reading all at the same time. AI fixes all that so you can be done in time for dinner. See how helpful it is? Ha!
August 6, 2026 — 9:10 AM
terribleminds says:
Ew. “The trouble writing a novel is all the cool parts. Wouldn’t it be better to skip the fun stuff and just lazily shit out somebody else’s book??”
August 6, 2026 — 9:31 AM
Kate says:
I’m fortunate in having recently retired from a software company where it was mandatory to use AI and report on its use every day. They had insane schedules and cut staff back because they said AI would do it all. It was incredibly toxic.
August 6, 2026 — 1:26 PM
Azuaron says:
“Oh, Sèphera, that’s not the point. The evil is GENERATIVE AI, not all the other AI!
Sure, sure. Let’s somehow separate one form from another.
Pick and choose your evil so that you can keep gatekeeping and let the oligarchs win as the little people continue to fight against each other.”
One of my biggest pet peeves in the WHOLE, ENTIRE WORLD right now is that AI does not exist. It is not a technology. It’s a marketing buzzword that covers a WHOLE HOST of different technologies. “AI”–seemingly according to the marketers–is just “a computer does statistics”; do you have any idea how many different computing technologies are based on STATISTICS? Come ON.
So, yeah, actually, we can pick and choose between “we wasted half the water of the Colorado River to make textual garbage” and “I turned a 2D photograph of Mars into a depth map on my PC in thirty seconds”. The first one is an obvious waste of resources, in a place where that resource SHOULD NOT be wasted. The second one gave me a cool piece of 3D art to hang on my wall basically for free, using existing tools that I have, with electricity I personally generated from the Sun.
Different things, it turns out, are different.
August 6, 2026 — 2:39 PM
annemerino says:
Excellent! Thank you!
August 6, 2026 — 8:09 PM
acflory says:
Well, bloody, said! Salutations from Australia.
By the way, I’ve recently discovered a couple of apps that may help us fight back against AI. Unfortunately, they are for graphics only at this point, but it’s a beginning. I haven’t been able to write a word [of fiction] since this whole AI debacle blew up. Now, at least, I have hope that someone somewhere will come up with a way to screw the AI trying to screw us.
For anyone interested, the graphics tools are called Glaze and Nightshade respectively. And no, I am not getting anything from either tool because they’re free. 😀
August 7, 2026 — 1:44 AM
Sarah Roark says:
(Part III to my reply to Chuck, sorry for the borked threading but your site is giving me technical issues with posting replies.)
August 7, 2026 — 9:47 PM
Sarah Roark says:
Okay, for some reason it’s just not letting me post my reply to Chuck’s part e), sorry about that Chuck. (Or you’re welcome, whichever.)
August 7, 2026 — 9:54 PM