The High Cost of Cheap Art
It seems impossible to avoid Artificial Intelligence in 2025. Specifically, generative AI: Programs that generate text, images, or music based on prompts. I know this isn’t exactly a novel topic, many people are against these new programs that corporations are desperately trying to push onto their platforms. I’m writing about it today because of how it specifically applies to me, Flaminia. It’s especially disappointing seeing this technology used by fellow creatives in the drag scene, and I hope I can encourage some of you to reconsider using it in the future.
When this tech started becoming accessible to the public, I was admittedly interested and entertained by it. Finally, I could know what it would look like if Donkey Kong attended the Nuremberg Trials, or see Kate Moss covered in actual Moss. It seemed like a harmless funny toy, and a kind of uncanny reflection of pop culture and the Internet. Maybe with some practical applications too, like artificially extending photo borders, or reducing grain. So, what’s wrong with it?
The biggest reason I won’t touch generative AI anymore is learning about the environmental impact. It uses a lot electricity to run, and then requires large amounts of fresh water to keep servers cool. It’s easy to not think about the material cost of prompts, but there are real computers in data centers using real power and water to do this. The only reason users aren’t paying now is because of major corporate subsidies (if only there was a way to make corporations subsidize cleaning up the pollution they produce, hmm). Newer models are only becoming less energy efficient as databases and algorithms grow larger and more complex. Even if this technology did become more energy efficient, it still poses a lot of ethical issues.
The way language learning models and image generators work (In an abstract sense, I do not have the time nor expertise to get into specifics) is by using extremely large datasets. When you ask ChatGPT a question, it isn’t answering you per se. It’s predicting what string of words to print out based on the prompt you provided. The data it uses for this is scraped from all over the internet and from trademarked work. Most of the artists and writers did not give permission for their works to be used this way. Some proponents of genAI argue that it’s taking inspiration the same way human artists observe each other. This implies thought and intent that these machines inherently lack. Processing words as data and churning out material based on that will always be plagiarism.
Despite art being fundamental to society, artists are treated as disposable. Art and entertainment is what makes life worth living. These industries generate countless amounts of money, but the creative people doing the work see far less of it than executives and IP holders. In my lifetime alone, there have been two major writers strikes to fight for better working conditions. Cultivating any creative skill takes years of study and has no guarantee of financial stability. With such a limited amount of work, it’s disappointing seeing potential opportunities get lost. It’s even more disappointing seeing other drag artists doing this, since we should understand the value of creative labor more… and we should have better taste than this.
I know this is subjective, but AI generated content sucks. Even when there are no apparent glitches (which there usually are) it always feels lifeless, dull, and lacking intent. Ok now disclaimer… I’m not trying to call anyone out specifically here… but I am going to give some examples. Love you all (except for the people I don’t, xoxo).
The most common use of AI I see from drag artists are trending photo filters. I’m not opposed to editing photos to suit a fantasy. I make most of my photos look like there’s Vaseline on an old camera. There’s no vision behind shallow trends though. One example I remember recently was queens (and a few kings) using AI to make their photos look like 90’s high school year book pictures. They were all eerily conventionally attractive. Still recognizable (even the ugly ones), but with slimmer jaws and airbrushed faces. Making a generically hotter version of yourself in clothes you did not pick and makeup you did not apply seems antithetical to drag as a form of radical self-expression. I think this usage would be relatively harmless (albeit tacky) if it wasn’t for the environmental impact.
The next most common use is posters. I get it, graphic design is hard, and new shows are happening constantly. But there’s a large overlap of drag artists and graphic designers, and many of them would love the chance to design flyers. If you can’t afford it, get creative. So much of the drag economy is built around bartering. We all have different skills, and we’re much more powerful when we use those together. I’ve done photo shoots in exchange for getting something sewn. I would totally design a poster in exchange for choreography help or something like that. If you can’t buy a design, and can’t barter for a design, just get a CC0 template online. A poster is a chance to say something about your show, and the only thing AI imagery says is that you didn’t care enough to make it properly.
Last year for a comedy pageant one of the contestants used AI to write some jokes… and it was obvious. There was a joke about me, Flaminia, lacking personality. Ironic considering it was written by something with no personality, and a choice that no real person would make. There are plenty of things that you could roast me, Flaminia, for; I definitely have a personality though, and a pretty eccentric one at that. Besides the subject matter not being applicable the delivery was stilted and off. ChatGPT has an “accent”; It sounds like a mix of corporation and forum poster. Even a bad joke would have been better than that word salad, and it would have been a better chance for that entertainer to grow.
The worst example of AI I ran into, which is the reason I decided to write this, was from a company I was scheduled to work with. Again, I’m not naming names. This isn’t a call out post. I know the people involved were acting in good faith. However, I want to talk about it because it’s a perfect microcosm of everything I hate about genAI. This event was not a traditional drag show (which is why I was interested in the first place). It started as an event for people to meet and was branching into queer mixers due to popular demand. I don’t have a problem with straight people running queer events, with the caveat that it’s being done thoughtfully and respectfully. ChatGPT and Stable Diffusion are neither.
The first red flag was the poster. Besides the clashing colors and bad composition I noticed more glaring issues. The title was nowhere to be found. Instead there was just a big “ATTENTION LGBTQIA+” in rainbow letters. It might as well have said “how do you do, fellow gays?”. The background had a multicultural queer group of friends that were also… not real. Their smiles were a bit too wide for their faces and there was some other uncanny quality to them. Something about that image felt particularly gross. It’s a similar kind of ick I feel from certain corporate pride material. Reducing our real community into a package to sell back to us.
The way the descriptive text talked about me, Flaminia, felt so condescending. Remember in my last article I mentioned how I felt like there’s a “default diva”? A drag queen with no personality outside of being a drag queen. Well, ChatGPT definitely knows that girl. Receiving an insult that didn’t fit was bad, but getting praise that didn’t fit felt so much worse. The prompt for the description was probably something like “write an event description for a singles event with a drag queen host”. I resent being called sassy, fierce, fabulous, and other generic phrases that make it obvious I’m being seen as a drag queen instead of as an individual. I’ve worked too hard doing things outside the box to be put back in one.
What finally made me pull the plug on working with this event was seeing the internal marketing. There was a line saying you could “meet someone who understands your point of view (and maybe even your pronouns)”. It’s the kind of cringey transphobia that you’d hear from a stale boomer comedian’s Netflix special (which is probably where the bot scraped it from). It didn’t get posted, and hopefully one of the human beings working on the marketing team would have noticed it and taken it out, but it still sucks. I wanted to make this work. I still like the premise of the event, and I hope the attendees have a great time. I just couldn’t be queer quality assurance and make new marketing material on top of what I already do; Not faster than just typing in a sentence, anyways.
2025 is simultaneously a renaissance and dark age for artists. We have new mediums to experiment with, and independent artists potentially have access to wider audiences than ever before. Outside of the control of major publishers. Yet, it’s easier than ever to erase artists too. To scrape their real thoughts, feelings, and ideas to pump out tons of nonsense for feeds. There’s financial motivation to bury real creativity under unfathomable waves of fake slop. If you’re just motivated by a bottom line, I probably can’t convince you to stop using generative AI. If you care about the planet, artists, and the integrity of your work, then please be conscious of what you put into the world.
