Editorial | Writing | Medium
I Click “Show Less” on AI
And you should too
On a recent call with investors, Hasbro CEO Chris Cocks claimed that most Dungeons & Dragons players he knows personally rely on derivative AI. This was either a complete lie or he doesn’t know any normal people, just tech-bros.
But it really doesn’t matter if his claim is true or not. The statement was for big investors who were probably blowing up his phone asking, “How are you integrating AI into EVERYTHING HASBRO SELLS?” Even if they don’t actually think AI-integration will help sales, they probably hope a big AI initiative will improve the stock price.
I’ve had to rewrite this piece at least twice, mainly because Medium’s rules on AI have been a bit of a moving target. Last year, Medium published what I thought was its new AI guidance. Later, I was informed that that wasn’t the final word on the subject and that there was an actual AI policy. Then, at some unknown point, that AI policy was updated. With any luck, that policy won’t change too much in the 15 minutes it takes me to reread the policy, re-edit the piece, and finally hit “Publish”.
Basically, Medium requires you to fess up to using AI near the top of the article. While it distinguishes between AI generation and AI assistance, they still want you to label AI-anything content. Otherwise the article won’t be distributed outside of your follower-base.
But I have NEVER seen such a label on ANY article on Medium. And about half the obviously AI-generated pictures are unlabeled too. And because Medium is crawling with AI evangelists, I refuse to believe the lack of labeling means nobody is publishing AI-written articles.
Note that this doesn’t apply to grammar-checkers, but Grammarly started branding itself as AI-powered about a year ago. Why does Grammarly not count? Probably because everyone uses it.
Up until last year, I used Hemmingway to help edit my writing. It didn’t check for grammar. It mostly pointed out hard-to-read sentences and passive statements, and that was about it. Then Hemingway implemented AI functionality and I cut that shit out. I didn’t want to help train another large language model.
That makes me — if you use Grammarly or any other AI-powered writing tool — better than you.
Derivative AI bugs the holy hell out of me, for a lot of reasons. For starters, AI tech gobbles up so much energy and water, tech companies are starting to quietly pull back on their green promises. Not that they were ever going to meet their net-zero goals, but now they’re not even going to pretend.
On top of the environmental nightmare, AI-generated text, music, and images are based on stolen works. Those creators were never compensated for their work. Yeah, you can train an LMM with only content you own, but that’s just a smokescreen for the rest of that shady industry.
A while back, in an act of sheer spite, I began clicking “Show Less” on every single Medium article with an AI-generated thumbnail. I don’t outright mute or block these authors. I hope that one day they see the light and stop feeding the Cylons. I save the mute option for writers who publish nothing but articles about how awesome AI is. And I don’t discriminate against writers who espouse AI’s potential life-saving functions. AI is for detecting cancer and heart disease. AI is not for making shite cartoons and shite essays and shite thumbnails and shite trailers for nonexistent Mad Max movies starring Donald Trump.
I also let AI-generated profile pics slide, although that may change soon. Some people want to brand themselves but don’t want to actually show their face. I sympathize. Then again, I use a glorified emoticon and I get by.
But there’s absolutely no reason to use AI images in a Medium article. Medium has a ton of pics from Upsplash. If you can’t find the perfect image, use your words instead.
A month into my Show Less crusade, I took the next step of calling for submissions about AI-generated media and media about AI. In retrospect, I should have focused only on the latter. Humans are story-driven animals, and the original Terminator movie can still scare the shit out of people.
Then again, our real-life AI apocalypse isn’t primarily driven by the military-industrial complex, like it was in Terminator and Battlestar Galactica. Instead, we can thank cheap bosses and greedy investors and tech bros for all these seven-fingered anime girls. Forget the “paperclip maximizer” problem. Most of us will be unemployed, homeless, and starving before “real” AI can properly destroy humanity.
I get why AI-bros exist. These tools make it possible to become an instant media mogul without having to hire ANY people. If you can afford pro- and enterprise-level AI tools and can drive them, you’d have a ready-made robot staff. A staff that does EXACTLY what you tell it to do, with no push-back and no opinion of its own. Combined with the slew of existing whatever-as-a-service tools, AI-bros can finally make their dream of being professional “Idea Guys” come true. And as AI penetrates more industries, these Idea Guys can enter more markets without any expertise or any employees. They just need the cash for the right apps.
And then there’s the nazi angle:
“I’m convinced that AI is *the* aesthetic of fascism in the 21st century a la Italian Futurism. “Art” without artists subordinated to political ends but (literally) untethered from reality… mechanically reproduced slop to promote apartheid and genocide[.]”
The Venn diagram of Trump supporters and AI enthusiasts isn’t quite a circle, but there’s more than a little overlap. Most of the politically-charged AI images on Xitter are pro-Trump or anti-Harris or generally Rightist, and I mute those accounts. Of course, there are a few Left-leaning AI images out there, and they piss me off too. No matter how much I agree with the message, I refuse to Like any of it. And if someone on the left posts AI crap often enough for me to notice, into the Mute pile they go.
On the subject of AI in roleplaying games, a snarky Bluesky user commented that…
“The only RPG that should use any sort of AI is Paranoia, and the AI in question should be absolutely barking insane.”
I’m sure he didn’t mean it literally, but I’m equally sure someone out there will do just that, using cleverness to excuse the fact that their automating what’s supposed to be a creative outlet. Now that derivative AI is starting to infect the TTRPG space, I’ve been driven to mute gamemasters who preach the gospel of ChatGPT.
But there is hope. An AI-curious TTRPG creator had previously dabbled with Abominable Intelligence but has seen the light and is remaking that project with human-made art. Fittingly, that project is a supplement to Cy_Borg, an artpunk-style game with explicit anti-corporate themes.
If you too have been led astray, know that it is never too late to come back to the light. Step away from the stochastic parrots. Come home.
