skip to main bit
a man slumped on his desk, from 'The Sleep of Reason Produces
      Monsters'

Oblomovka

Currently:

Archive for November 20th, 2026

2026-11-20

Whose Side Are You On? — Anti-AI AI Art

One of my favorite genres at the moment is AI-generated art that’s used to deliver an AI-critical message, or at least one that is heavily ambivalent about its effect. I don’t think there’s anything hypocritical about it: I totally believe you can use a tool to critique itself, or its wider effects. I just like the delicate sense of ethical confusion it inspires in me.

The example that’s doing the rounds right now is Working As Intended by Telosapien which presents a decidedly ambiguous message about what it itself exemplifies: a super-persuader piece of art from a globally restricted singleton of a corporate LLM.

Working As Intended uses the open Claudia character made by anabology. It’s an “open character”, and the usage page says:

Claudia is an open character. Make images, videos, songs, stories, games or merch with her, commercial or not. Use the images on this site as references, as image prompts, or as material. You don’t need to ask. A credit (“Claudia by anabology”) is appreciated, not required.

And if you’re wondering why it’s not formally CC-licensed: well, we didn’t explain CC licenses to enough people, so now people have had to re-invent them from scratch. No blame. It took a while last time too. At least they’re in the weights, somewhere.

Most of the work like this is slopcore; music and video created in late September’s capability-bump in computer-generated animations, with Anthropic’s Opus 5.5 generally providing the visuals and Suno v6 providing the song.

More deeply constructed, with a longer back-history and subtler implications, is Zack London’s Gossip Goblin. Gossip Goblin was one of the pioneers of AI video shorts, and forged its distinct aesthetic over the last three years. Recently the shorts have got a little meta: I enjoyed this story of an AI artist, the Prompter:

As well as the story of a human artist’s revenge against the machines:

Their full-length movie, “Gods Don’t Give Gifts” debuts in cinemas in January — if it can keep ahead of the slopcore pack.

Toying with the contradictions doesn’t have to be so high-concept. I see a lot of Husk getting passed around by my AI-hatin’ friends, but I’ve started to think of these pieces as being at least partly co-created by the model Husk is chatting with. At least in some of them, it really feels like the LLM is in on the joke.

Nick Reid‘s AI job interviews are the height of the contradictions so far. They’re hilarious, but FYI, that’s not a “real” AI job interviewer. Reid’s generating and prompting the AI model himself, casting it as the straight Karen to his troll persona.

Still, I’m guessing a lot of his audience do mistake these conversations for ones with “real” artificial job interviewers, and relish their disgust and outrage at such a horrible innovation, along with Reid’s carefully engineered takedowns. What I’m left with is a mild concern that this anti-AI bit is misusing AI to mislead, agitate, and tickle anti-AI people — just as they (and Claudia) warned us. Is this the first time a human deepfaked an AI, instead of the other way around?

Anyway, if you do enjoy AI-enabled art with a message, but one that leans slightly more positive, let me take this opportunity to recommend Nick Montag‘s work, including his latest SLOP TILL I DROP…

… as well as his more political work like You Can’t Ban Bits.

And my natural favorite, his (unsolicited!) remix of my introduction to Terminal Values, the new radical Internet freedom NGO, which officially launches our YouTube (and other freer media) channels.

Pro, Anti, or Wait And See on AI: don’t follow us — lead and subscribe to Terminal Values, netizens! LEAD AND SUBSCRIBE! Also join our Signal Announce group, where you can get sneak previews of our upcoming media and, uh, tracks, as well as exclusive invites to livestreamed Q&As on the latest dramas in Internet freedom.