Thread by @gralefrit

Biblendum

#nichejokes

Ah. Bollocks. Have found out that this is a fucking Midjourney AI image. I mean, wrecking humans’ fun at daft caption games was not something I saw coming, but there you go. AI really is the Covid 19 of comedy.

“We’ve made an image of something that didn’t happen, which can escape the context of its maybe originally amusing text prompt, to be recaptioned by everyone the same way, with the same joke, about its incongruity, which doesn’t exist, because it never happened.” Oh. Helpful. Ta.

Whatever joke you apply to this image, it’s only satisfying because you are forcing sense onto an image which initially seems improbable. If the image was designed to be improbable, then the humour dies.

It’s like captioning a clip of Basil Fawlty thrashing his Austin Maxi with a line about “Jeez! This guy seems crazy angry!!!!”

Once AI images escape their context, all attempts to joke about them become that unwelcome “extra” gag people tweet under a joke that was already finished satisfactorily, merely stating the joke again, more weakly. Ooh. This has made me proper vexed, this has.

Also: if you’re the one pointing out the image people are playing caption games with isn’t real, you become the party pooping killjoy. Even though you’re completely right to point out that the comedy is weak. It makes everyone’s position around the joke awkward, unsafe. Awful.

I was already wary of AI’s risk to news and comment - that a fake image that seemed to tell a story would always escape and spread. But spoiling one of the simplest forms of shared comedy wasn’t something I saw coming. This stuff really poisons the well, for zero gain.

Anyway, I’ve deleted any other jokes I made about this image. Which felt shitty. The information that it was a Midjourney image was on one of those threads. Maybe it isn’t, but it seemed legit. And now that’s a waste of my brain and time and feelings. I don’t get what we gain.

Chris Morris said that part of the craft of doing cut-ups and fake interviews and so on, was that the audience had to see that it wasn’t real. You had to leave in some phuts and clicks and audible splices. Or there was no way in to the joke. This AI shit steamrollers that craft.

It’s an area of comedy that always makes me uncomfortable; that the purpose of a prank is to seamlessly fool someone, for it to pass as real. That’s not a joke. That’s bullying, or superiority. It’s a trick that you’re not meant to detect. But the fun is showing your working.

Humour is a social act. The key word in “sharing a joke” is “sharing”. It’s fundamentally human. And the robots aren’t understanding that.

“Loads on why jokes do and don’t work in my forthcoming book” tweet here.

unbound.com/books/comedy-b…

Basically, jokes (often) resolve incongruity by providing an alternate pattern or prototype by which nonsense is turned into sense on second examination. In my caption, a pun allowed the Pope to also be the Michelin Man. Explaining the mad coat.

But if there was no incongruity, and the pope never wore a mad coat, my pun is not a new mental model that helps return our internal error-message red light to safe green. The light is already green; the incongruity never existed. So no joke is required to render it safe.

I’m not going to try and condense a 90k word book into a dozen tweets, but making a gag based on an amusing image won’t work if the image was created for the purpose of seeming amusing. Or it will, until the moment you find out it wasn’t real. At which point you’ll feel sad.

TLDR: I have seen the future of comedy and it is a robot that has been programmed to think it’s Alison Jackson.