Kylie Robison, reporting for The Verge:

The trend kicked off pretty wholesomely. Couples transformed portraits, pet owners generated cartoonish cats, and many people are busily Ghibli-fying their families (I’ve stuck to selfies, not wanting to share with OpenAI my siblings’ likenesses). It’s an AI-generated version of the human-drawn art commissions people offer on Etsy — you and your loved ones, in the style of your favorite anime.

It didn’t take long for the trend to go full chaos mode. Nothing is sacred: the Twin Towers on 9/11, JFK’s assassination, Nvidia CEO Jensen Huang signing a woman’s chest, President Donald Trump’s infamous group photo with Jeffrey Epstein, and even OpenAI CEO Sam Altman’s congressional testimony have all been reimagined with that distinctive Ghibli whimsy (it’s not clear whether these users transformed uploaded images, or prompted the system to copy them). Altman has played into the trend too — he even changed his X profile picture into a Ghibli rendering of himself and encouraged his followers to make him a new one.

I’ve expressed disgust at artificial intelligence-generated images before, most notably in my Apple Intelligence article last June, so when people started posting stills from “Severance” styled like art from Studio Ghibli, the famous Japanese animation studio, I felt some mild discomfort but most left the dust to settle. And it really did settle — the craze ended less than 24 hours after ChatGPT 4o’s new image generation tool rolled out to paid subscribers because OpenAI pulled the plug on generating characters resembling copyrighted work overnight, at least for free users.

But nothing really stuck out as truly repulsive to me. It didn’t even seem worthy to write about. That was until the White House posted an image of a woman — seemingly a fentanyl dealer — being deported by Immigration and Customs Enforcement in the style of Studio Ghibli art, apparently created using GPT-4o. Detestable. The post is still up on Elon Musk’s 4chan knockoff, X, and I highly doubt it’ll be deleted after the same account posted an ASMR — autonomous sensory meridian response; a quiet piece of content meant to be relaxing — video of migrants being loaded on planes and deported a few months ago. But while that video was also vile, it didn’t strike me the same way the AI-generated image did.

I’ve been trying to piece together why I was so viscerally taken aback by the image. I know the White House. I know about the detestable Nazis who work there. Nothing they do surprises me even in the slightest. If Stephen Miller, the administration’s “border czar,” started belting out the N-word tomorrow, I wouldn’t even bat an eye. (For clarification, that doesn’t mean I agree with them — it’s that I wouldn’t be shocked.) Slowly, the realization kicked in: I’m disgusted that OpenAI, a company made to create AI that benefits all of humanity, let this slide. It’s impossible to ask ChatGPT to generate something as harmless as erotica or a violent fictional story, but it’s OK to create images that depict humans as livestock? How does this technology benefit humanity?

Worse of all, the image was generated in the style of real artists. It models the work of a real studio. OpenAI is profiting off the Adolf Hitler fanboy club’s wet dreams about beating up migrants while stealing the work of real artists. This dehumanizing, animalistic post looks like an endorsement of the Trump administration by Studio Ghibli itself, but it isn’t. It’s far from one. It’s an endorsement by Altman and his cadre of Silicon Valley extremists. We’ve reached a new low in the human race where it’s acceptable to steal a studio’s hard work and use it to depict humans like animals, all while making billions of dollars in revenue. Where is the “open” in OpenAI? How does this adhere to the company’s mission statement? From OpenAI’s charter:

OpenAI’s mission is to ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity. We will attempt to directly build safe and beneficial AGI, but will also consider our mission fulfilled if our work aids others to achieve this outcome.

Again, how does creating Nazi propaganda posters benefit humanity? Who is this benefiting? How does stealing the work of artists benefit humanity, let alone all of it? How does blatant copyright infringement get us any closer to AGI? Nobody’s answering these questions. Nobody’s answering how ChatGPT is perfectly fine generating images of people being dehumanized. There’s no safety at OpenAI anymore — it’s just a group of low-life grifters with no spine. Need proof? Why is the company’s head encouraging users to generate more copyright infringement slop through his product? Even Musk, a person truly a waste of this planet’s natural resources, shut down the Twitter verification system temporarily in November 2022 after people made images of Mario giving the middle finger. But Altman has no shame, and he’ll do anything to get in Musk’s position for the money.

It was a mistake to lift the safety nets over ChatGPT’s image generation. The new version of GPT-4o has nearly none of the guardrails first introduced with DALL-E, OpenAI’s first image generator. That’s because when DALL-E first came to market, OpenAI had morals. Joe Biden was still president. Altman hadn’t been fired and rehired, causing a revolt in the company and boosting his ego beyond proportion. GPT-4o now generates indiscernible images of people, engages in blatant copyright infringement, and has no regard for humanity’s benefit whatsoever. It’s just like Grok’s image generator, only used by hundreds of millions of people around the globe. Forget the purported dangers of generative artificial intelligence, which I’m still skeptical about: this is Step 1 in AI accelerationists’ plan to devalue humanity, creative expression, and morality.

AI won’t revolt against humans or take everyone’s jobs. No stupid computer will ever steal a single Studio Ghibli artist’s job. Not now, not ever. This is not the movie “Her.” AI, however, will make the world a deeply immoral place. It’s the modern equivalent of sea pirates, where the laws are controlled by self-proclaimed monarchs, the courts don’t exist, and the oligarchy rules the poor schmucks doing the work. This is happening in the United States right now, and there’s nobody to stop it. Pointless slop image generators are the beginning of an era of moral bankruptcy.

And while the moral bankruptcy certainly lies in part within the people who use AI image generators for nefarious reasons, like the White House, it’s even more the fault of the AI companies themselves for failing to create safeguards. We have no meaningful AI regulation — not in the United States or the world at large — so it’s up to the AI industry to self-regulate. But no business on planet Earth regulates itself, no matter how humane or ethical it might purport to be. It’s akin to school shootings in America: the gun lobby will never advocate for a law that bans assault rifles because that would be against its bottom line. Shooting up an elementary school is already illegal, and so is copying another artist’s style and selling it for $20 a month. Making the user’s action illegal won’t solve the problem because it’s already illegal and nobody cares.

Regulate the AI image generators before it’s too late.