How Tom Short learned the skills – and limitations – of AI

If you'd told me ten years ago that one day I'd be taking a show to the Edinburgh Fringe written with the help of artificial intelligence, I'd have assumed you were pitching a dystopian sci-fi film.

If you'd told me five years ago, I'd have probably argued with you.

If you'd told me two years ago, I'd have asked whether the robot was available to help with my tax return.

Yet at the Fringe, I’m performing as show called I Get ChatGPT To Write My Edinburgh Show.

The strange thing is that it didn't begin as a serious artistic project. It started as a joke.

As a comedian and lecturer in comedy writing, I'd spent years teaching people how jokes are constructed. I'd talked about character, audience expectations, tension, release, rhythm and surprise. Then suddenly this new technology appeared that claimed it could write comedy.

Most comedians reacted in one of two ways. They either dismissed it completely or became convinced it was about to replace us all.

My reaction was simpler.

I wondered what would happen if I put it on stage.

The original idea was little more than a gimmick. I'd ask ChatGPT for comedy material, perform the results and let the audience watch the experiment unfold in real time.

What I didn't expect was how quickly the project stopped being about artificial intelligence.

The more I worked on the show, the more I found myself asking uncomfortable questions about why I was making it in the first place.

Why am I so fascinated by technology?

Why am I interested in the idea of a machine creating art?

Why do performers spend so much time chasing approval from strangers?

Why does the internet seem to encourage all of us to become our own marketing departments?

And perhaps most importantly: what are we actually afraid of losing?

The easy version of this story is that comedians are worried about robots taking their jobs.

The more complicated version is that many of us are already competing with algorithms every day. We perform in a world where attention has become a currency. Success increasingly depends on visibility. The pressure to create content never stops.

In that environment, AI feels less like an alien invasion and more like the next logical step.

As I continued developing the show, I realised that ChatGPT wasn't actually the subject.

It was a mirror.

Every prompt revealed something about the person writing it. Every interaction forced me to think about my own creative process. The machine wasn't exposing the future of comedy nearly as much as it was exposing my relationship with comedy.

That's when the show became personal.

Underneath all the jokes about artificial intelligence is a show about validation, ambition and the strange things performers do in the hope of being noticed.

It's about growing up working class and choosing a career that many people don't fully understand.

It's about spending years trying to build an audience.

It's about the temptation to believe that the next piece of technology, the next platform or the next opportunity will finally provide the acceptance you've been searching for.

The irony is that the more time I spent asking ChatGPT questions, the more the answers pointed back towards me.

The machine couldn't tell me whether I was funny.

It couldn't tell me whether the show would succeed.

It certainly couldn't stand on stage and perform it.

In the end, that's the part of comedy that remains stubbornly human.

A joke only really exists when another person hears it.

Which means that despite all the headlines, all the predictions and all the panic, the most important part of comedy hasn't changed at all.

It's still just people in a room, trying to connect with one another.

The robot is merely helping me talk about it.

• Tom Short: I Get ChatGPT To Write My Edinburgh Show is on at Hoots @ Nicolson Square at 22:15

Published: 17 Aug 2026