Last week, I was scrolling through my local queer community Facebook group when I came across a post from a barber advertising her services to the queer community in my city. The accompanying graphic had all the telltale signs of AI generation, from some wonky text alignment to that disgusting yellowish-sepia tone and rounded graphics that seem omnipresent in many AI graphics.
There was only one comment on the post, from another well-meaning user advising the barber that using generative AI in her promotion might turn people off of her services. Ten minutes later, I went back to look and the post was gone. Evidently, a little bit of public shame was enough for this user to reconsider—at least publicly—using the technology in that way.
I’ve been thinking about this exchange as discourse continues to build around the new film Stop! That! Train! and its apparent use of AI-generated background imagery. Starring a pile of Drag Race alums alongside RuPaul herself, the film is pitched as a campy and fun take on disaster movies, as various drag queens star as stewardesses on a runaway train. But shortly after its premiere at NewFest in New York last month, fan speculation emerged that several of the film’s effects were made using generative AI. One Letterboxd review from visual effects artist Gloria Cook went viral, where she claimed it was one of the most “conspicuous uses of AI” she had ever seen. This spurred a wave of fan backlash. Many pointed to the film’s credits as evidence: Acme AI is top-billed source of visual effects for the film, and the lack of credits for other visual effects artists.
Adam Shankman, the film’s director, initially denied that generative AI was used to “conceive shots” in a mealy-mouthed social media statement last week.
“Every shot in Stop! That! Train! was made by human hands! It’s come to my attention that there is some on-line speculation that Stop! That! Train! is full of fully generative AI shots and I’m here to tell you this is patently not true. There are a sum total of ZERO shots conceived by AI in the movie,” Shankman wrote.
But to paraphrase Dakota Johnson on Ellen, “… that’s not the truth, Adam.” In an interview with Russ Martin published here at Xtra today (conducted before Shankman’s social media statement), Shankman seemingly contradicted himself, admitting that there actually is a fair bit of AI in the film.
“Some of the windows were AI. There was only one AI shot or two AI shots of the train, the rest are CG. Most of the CG is the train, then out the windows is mostly AI. But the Florida station, all that, that’s CG. And different licensed imagery.”
That lines up with the evidence compiled by various Reddit sleuths over the past few weeks, particularly around the titular train’s windows changing from shot to shot. Cook pointed out in a follow-up social media thread that this would not be the case if a real computer-generated model, rather than generative AI, was used.
A generous reading of Shankman’s contradictory comment suggests there is a degree of semantics at play here around what constitutes AI. In his initial denial, Shankman was talking about shots “conceived by AI” (perhaps alluding to part of his job as director), and it seems likely that he’s telling the truth that he never prompted AI to make a shot for the film. But fans are outraged about AI-generated visuals in general, and that is clearly something that Stop! That! Train! includes, according to Shankman’s interview with Xtra.
This isn’t the first time a production has gotten into hot water around the use of generative AI. Back in 2021, the film Roadrunner: A Film about Anthony Bourdain, used AI to read several of Bourdain’s emails posthumously. Netflix faced backlash in 2024 for allegedly using AI-generated images in their documentary What Jennifer Did, though the film’s director issued a similarly vague denial as Shankman saying the film doesn’t use AI but does use “various tools” to alter images shown onscreen. This isn’t even the first time a RuPaul production is in hot water over generative AI visuals—an April episode of Drag Race featured images of the season’s final four that seemed AI generated.
This sparks a moral dilemma for those of us who feel opposed to generative AI, but want to support a theatrically released, gay-as-hell movie during Pride Month. Should we just accept that we don’t have to like it, still watch the movie and move on? Should we boycott out of moral or ethical concerns, as Cook and other fans on social media are suggesting?
These days, everyone has a line that they draw personally when it comes to AI. For me, knowing that something involves AI-made imagery, writing or music makes me not want to engage with it. That goes for the barber in the Facebook group, and it goes for all of that nonsense swirling around “AI actress” Tilly Norwood or AI “music.” And it goes for movies like Stop! That! Train!, where I now know it played a role.
There are plenty of reasons for my personal opposition to generative AI. It carries a huge environmental toll, prevents some from paying and supporting real creative people and generally degrades truth and reality in our society. On top of that, everything it makes is straight up worse. And maybe it’s too high a moral soapbox to stand on, but I lose respect for people who use generative AI to make creative work because it’s clear they don’t respect the labour that goes into making art. And that includes visual effects artists who make visuals of a gay runaway train in a campy spoof movie.
I can control my own life on a small scale—I have no interest in using ChatGPT day to day, and only use programs with generative AI tools built in where there is no other alternative, like Canva or Google—but it is getting harder to avoid generative AI in media. Sure, I can choose not to go to that barber who made the AI promotional post, or avoid that restaurant in my neighbourhood plastered in horrifically uncanny AI food images. But I find it challenging to adopt a personal hard-line stance across the board, because it truly is everywhere, especially in subtle ways I might not even notice. Stop! That! Train! just happens to be relatively obvious.
Boosters of generative AI technology in media—or even seemingly pragmatic late adopters of it—often talk with an air of inevitability. They claim that skeptics of the technology just don’t see the light, or we’re not interested in innovation. To them, generative AI is inevitable, like the internet or social media once was. Even the great filmmaker Martin Scorsese is on board, investing his money in an AI image-generation company and telling the New York Times last week, “Remember, cinema is a young medium, only around 125 years old, so we have to be open to how it can evolve.”
At the same time, artists of all stripes—see, the recent writers’ and actors’ strikes in Hollywood—have been clear that the surge in use of generative AI presents an existential threat to human creation. While some wonky train windows might seem harmless at this point, there is a very fast-moving slippery slope that brings us to other kinds of AI content. Just ask the thousands of journalists who’ve been laid off as the news media industry increasingly pumps out AI-generated writing, or the animators who have lost funding to make their work in favour of Amazon making an “AI creators fund.”
At the end of the day, I can’t tell you to boycott this film, even though I kind of want to. It would be hypocritical because we all have to encounter generative AI out in the world these days whether we like it or not. But what I can do is encourage you to do—whether you’re going to see Stop! That! Train! or not—is to also seek out real artists making real things. Calls for boycotts and public shame can be effective and emotionally cathartic, but so is supporting the creatives making an intentional effort to not include AI in their work. Credit where credit is due: when tasked with including a seemingly AI-generated meme in the recent The Devil Wears Prada 2, the filmmakers actually hired a real artist to make a painting that looked like AI. And Kane Parsons, the wunderkind director behind horror smash Backrooms, said he didn’t use any in his film and said AI leads to “cultural and economic rot.”
In the case of Stop! That! Train!, you can also go to your local drag show and tip your queens. Put your valuation of human-made art into action. Give a dollar (or better yet, $20 or $50!) to a real queer person making real queer art. If we want to survive the runaway train that is AI in media, the best way to start is by sticking together and standing up for what we believe in.
Why you can trust Xtra
