AI Comedy Marketing

David Schneider: why AI is as useless as a marzipan dildo when it comes to comedy

Author

By David Schneider, Founder

July 20, 2023 | 12 min read

David Schneider has written a lot of good jokes. For an experiment, the writer, director and founder of creative social agency That Lot tried to do so using AI. And, well... he’s not out of a job yet.

Marzipan Dildo

We do not need ChatGPT to tell you that AI is everywhere. The only thing growing faster than AI is articles and think-pieces about AI – such as this one [and all of these]. The technology is improving so quickly that many worry we’ve already reached Terminator minus four months.

There’s no doubt we could be living through a revolution as big, if not bigger, than the internet. And when you think that it was the internet that brought us Susan Boyle’s team tweeting the hashtag #susanalbumparty, you get a sense of just how major this could be.

How worried should we creatives be about AI coming over here, taking our jobs and generally enslaving us? Take humor, for instance.

Here’s a joke written by a human, @0point5twins on Twitter:

“Describe yourself in three words”

“Lazy”

And here’s what ChatGPT came up with when asked to make a joke about laziness:

“Why did the lazy person become a mathematician?... because they could always count on doing the minimum.”

Now, that has the rhythm of the joke, it depends on a play on words like many jokes, sure, but good luck trying that out to a group of stag nights at the Comedy Store.

Even when you push ChatGPT with further prompts, the best it can do now is somewhere between Pound Shop Christmas crackers and mid-table dad jokes (apologies to my fellow mid-table dads out there).

When it comes to actual creativity, AI struggles to climb beyond base camp. Asking it to come up with copy for an image of Big Ben on the Adobe Instagram account, even when prompting it carefully on tone of voice and audience etc, produced options that ranged from the meh to the For-God’s-Sake-Shoot-Me-Now. Jewels like ‘A timeless icon of the London skyline’ or ‘Big Ben never looked so good!'.

Not that this is without its use. As Tom Roach of Jellyfish Global tweeted [and explained in The Drum], if AI’s come up with it, you know you can dismiss any similar thoughts you have as too obvious. It can also be helpful as a thought starter or to get around writer’s block.

One of its offerings on Big Ben for Adobe was:

“An inspiring symbol of creativity and innovation.”

It’s not great but you might notice that the statement could be applied to Big Ben and Adobe. This might then encourage a bit more noodling about commonalities to take you to something like:

‘Creativity and Time. Name a more iconic duo.’

Before bringing you to:

‘Best gift you can give a creative? No, not donuts. Time’

Or, punchier:

‘Best gift you can give a creative? Time’

AI can start the ball rolling before humans pick it up and run with it, thus inventing copywriting rugby (other, better metaphors are available).

A-Ha(i): could AI write Alan Partridge?

As one of the original writer-performers on Alan Partridge, I was honored to be present when Alan was born fully formed, in sports casual Pringle top and relaxed beige slacks. When throwing around ideas for his biography, one of the questions was: where does he live? Milton Keynes was an early suggestion but quickly rejected as being too ‘route one’. Eventually, we decided on Norwich – beautifully specific, with that whiff of Partidgean little England and that twist of surprise that good comedy requires.

AI is at the Milton Keynes phase at the moment. At best it’s a promising junior working for you, who’s great at sourcing information (as long as you know some of that information may be wrong and/or entirely made up) and who sometimes gets confused about the number of fingers a person should have in an image. It will get better. When it comes to humor, I can already see a huge improvement.

Here is its offering on laziness from a few weeks ago [warning: may cause your soul to pack it all in and leave your body]:

“Why did the lazy person become a gardener?”

“Because they heard it was a great way to ‘re-leaf’ stress while ‘sprouting’ excuses for not doing anything else!”

There are rules AI can learn, for instance about the need to surprise and the joy of specificity, pushing it from a more obvious to a more original choice, from “as useless as a chocolate teapot” to The Thick of It’s more robust: “as useless as a marzipan dildo”.

Useful as a marzipan

But will it ever understand why “marzipan” is funny? Not just the beauty of its specificity, but that it’s slightly old-fashioned, a concept hidden in some distant larder of our minds which we love re-discovering, that it’s vaguely associated with childhood memories and therefore even wronger when placed next to “dildo”, that the word itself sounds funny?

How can you break down all those (marzipan) layers for a machine? Surely our jobs are safe.

Suggested newsletters for you

Daily Briefing

Daily

Catch up on the most important stories of the day, curated by our editorial team.

Ads of the Week

Wednesday

See the best ads of the last week - all in one place.

The Drum Insider

Once a month

Learn how to pitch to our editors and get published on The Drum.

I can see how AI could (mass-)produce Milton Keynes content. But for that extra marzipan twist of brilliance, surely (do I sound desperate?) you need a human. The best creative has to stand out from everything that’s come before it; it has to be able to break the rules not just learn and follow them.

Fleabag was already breaking the rules to an extent when its lead talked to camera. But when in season two (spoiler alert), Hot Priest asked her who she was talking to, thus breaking the breaking of the fourth wall, a sort of rule-breaking squared occurred which also packed an incredible emotional punch (he really, really sees her). Could AI ever make such a surprising move?

Forging previously unseen, unexpected twists and links is what makes great creative (see this clip of Seinfeld spotting that both golf commentators and people talking about tipping talk in a whisper). It’s hard to be unpredictable if everything is mathematically predictable. Creativity needs a bit of chaos, a bit of grist in the mill, and a bit of rule-breaking if it’s to rise above the mundane.

So much of it comes down to instinct, paired with a knowledge of your audience. Ending a social media post for a particular Gen Z audience with a full stop is akin to donning a baseball cap and shouting “How do you do, fellow kids” (same, I imagine, with using the word “akin”).

‘AI doesn’t know what it’s like to stutter’

Similarly, having written comedy since floppy disks were floppy, I’ve seen how what constitutes a funny number varies across the years and according to audiences.

Sometimes an odd number is funnier, sometimes even, sometimes a round one. I reckon right now “a lot of people worry we’ve already reached Terminator minus four months” is funnier than “we’ve already reached Terminator minus three months”. There’s a precision to “four” that “three” lacks but above all, I have an innate sense (ie one that can’t be reduced to data) that, at the moment, even numbers are in. For a while, the number three (written as ‘three’ not ‘3’) was funny for a niche Gen Z audience due to this meme; same for the letter ‘h’, I’m reliably told by one of That Lot’s Gen Z creatives, Josh Markey, who may, of course, be lying to make a borderline Boomer look bad in this article [editor’s note: we had to fact-check this].

I’m no expert (not that it’s stopped me and thousands of others writing articles like this) but won’t AI always lag behind? Our jobs are, surely, safe. Surely.

Instinct is unlearnable. Watching Paul McCartney coming up with Get Back in the Beatles documentary was like watching a man being touched by the hand of God, even to this atheist. And then the relentless working and reworking, honing and experimenting, as if subject to unseen, inexplicable self-prompts until Macca could feel he had it, that it worked, that it was the shape it was always meant to be, like Rodin seeing his finished statue in a hunk of rock. This seems unteachable, unlearnable. An AI can’t get goosebumps, whether at its own work or someone else’s.

Above all, an AI can’t have empathy.

It was never bullied at school; felt different because it was part of a minority; or accidentally called its teacher “mummy” as I once did which I know isn’t that unusual... except I did it at university. As a placard held by the writer David Seidler said on a recent Writers Guild of America picket line: “I wrote The King’s Speech. AI doesn’t know what it’s like to stutter.”

Ultimately, if there is a war against the robots, our shame may be our superpower.

Obviously “Will we creatives still have a job?” is not quite as significant a question as “Will AI mean the extinction of all human life as we know it”, and it may well be that my optimism will prove as useless as a marzipan dildo.

But for now, as long as agencies continue to want excellence rather than competence, brilliance rather than mediocrity, and awards rather than completed content calendars, creatives are safe. At least for another week.

AI Comedy Marketing

More from AI

View all

Trending

Industry insights

View all
Add your own content +