This is the true story of CAI, the robot who made advertising

When you work in advertising, your job is to create messages that stand out and make an impact to create brand preferences among consumers. The message is always the same. "Buy me, I'm the best," says the ad.


CAI interface

When you are a creative in advertising, your job is to find new ways of saying this. We are here to sell things, everybody knows that. The aim is to do it with a wink of the eye that says: "hey, I know you know I am trying to flog you something, but look, I found a new way of doing it and you'll think its pretty cool." We’re just a wink factory really.

Today there are dozens of advertising festivals around the world that reward these new ways of saying the same thing, these new forms of rhetoric. With Grand Prix, golds and silvers, these chosen ideas strut the Croisette, the covers of professional magazines and are talked about on blogs. And when we see these campaigns, we say to ourselves that advertising still sparkles and is still able to reinvent itself in different forms and different media, to keep surprising us. It's quite exciting to see all that talent on display. It is an engine and an example not only for those who make the advertisements, but also for those who watch them and who commission them. But just between you and me, it’s more the lion that hides the savannah.

CAI ad

For everything that is being rewarded as the cream of the crop, this "new way" to sell us things is only the very tip of a huge wave, a tsunami of campaigns unfurling on people around the world. One only has to have been on a jury in one of the festivals to realise that what is rewarded only represents maybe 0.00001% of global production.
 And the rest? Well... to be honest it’s not really all that great, it's not really new. You just have to open a newspaper or turn on a TV to see too many repetitious, clichéd, mindless, condescending messages. It is counter-productive for our industry insofar as consumers become increasingly advertising-averse and our customers more wary of the effectiveness of campaigns.

It’s as if there isn’t one discourse per brand but a single formatted advertising language, a sort of “Esperanto” understood by everyone, a bland, unimaginative language of few words that tells everyone “pay attention, this is advertising, this is how to speak in advertising.”

CAI ad

It was this rather disturbing observation that gave the idea to push this paradox to its limit. Since advertising all too often resembles a mediocre rehashing of things that already exist, why not try to create a machine that would do it in our place?

And so the idea of CAI (Creative Artificial Intelligence) was born – a software robot that immediately and infinitely creates simplistic and non-differentiated advertising. CAI represents nine months of relentless work.

CAI is first and foremost the product of a huge amount of planning work. All possible, imaginable brand promises by product category were compiled. Next, CAI was loaded with thousands of visuals, bits of copy, and dozens of typical page layouts. So when you ask CAI to work, she can randomly generate around 200,000 ads.

CAI ad

Let’s get to the specifics. You type in your category product (e.g Junk Food) and then your product (Pizza). Next you have to choose your target audience, the aim of your campaign (create awareness, launch a product, etc.) and you type in your chosen brand name. Following this, CAI will generate your product pack-shot. CAI will then propose all the possible brand promises associated with your product (in our example – crunchy, genuine, homemade etc.). Once this is done (all in just one minute, exactly), CAI will present you with her copy strategy. If you are happy with it, CAI will begin her creative process (20 seconds) and offer you three possible print ads. Once you have made your choice, you can even see your future campaign in situ. Thank you CAI.

It started out as an intellectual game became more and more alarming as it progressed in the development of our robot. It was expected to create a clumsy, rather grotesque machine that would be systematically way off the mark, in a comical way. Unfortunately, this is not the case. CAI produces something that is a caricature, but that very often by some random diabolical grace, reminds us of an ad we have already seen on a street corner or on the page of a magazine.

CAI creates immediately and quasi-infinitely something that resembles advertising but that fundamentally isn’t, in the sense that it lacks essential qualities: novelty, inventiveness and the unexpected. In short, anything that only a human being is capable of producing. CAI is a fascinating but dangerous machine because it synthesizes the nemesis of our creative profession: standardised or formatted thinking, call it what you want.

In this sense, it is our responsibility to cultivate the differences between our agencies, to encourage our planners and our creatives to always strive to go further, to not be content with regurgitating what has already been seen on the web or elsewhere. This is what creates the value of our thought production. It is what we owe our clients and what they have the absolute right to demand of us.

Such is the true story of CAI, the first and, I hope the last, robot who made advertising.

Special thanks to my team:
Élodie Andurand
Héloïse Hooton
Clarisse Lacarrau
Claire Maoui
Xavier Royaux
Abder Zeghoud

This article was first published by Contagious Magazine # 22 [pdf link] by Stéphane Xibérras.

15 comments

Jaap Grolleman's picture
Jaap Grolleman
7000 pencils

As Luke Sullivan said: “Creativity is like washing a pig. It’s messy. It has no rules. No clear beginning, middle or end. It’s kind of a pain in the ass, and when you’re done, you’re not sure if the pig is really clean or even why you were washing a pig in the first place.”

I doubt a computer can ever take that over. The computers, as I know them, only work on logic and on what humans programmed them on, and never create an unexpected solution.

Also, when I see the above ads, I don't really fear for my job.

ivan's picture
ivan

Let's hope what you're saying is true. But to be honest randomness can indeed be programmed into an application. It's a matter of taming this randomness to generate meaningful results. Perhaps in the future it will be common for creatives to use such apps to help them generate ideas or organize thoughts, but creatives will be there to monitor and control the process for a very long time.

Jaap Grolleman's picture
Jaap Grolleman
7000 pencils

To call creative ideas random is a bit of an insult to our industry in my opinion hehe. But yeah I get what you're saying. I don't fear it though. Bring it on.

ivan's picture
ivan

I didn't say creative ideas=random ideas. I just said that there can be a random element built into it that can help create unexpected outcomes. :)

Guest's picture
Guest

If it's pre-programmed the 'idea' cannot be random, only the sequence of its combination. Real-world creativity is often most exciting when serendipity gets involved. You can't program for that.

ivan's picture
ivan

I'm not convinced about that. But I think we should not need to argue. Too much emotion is involved here. :)

observant vicky's picture
observant vicky
194 pencils

In some stages may be you are right but I must say a computer can't replace a human, in terms of creativity.....
for me creativity is observation of real life and real people and on this platform a robot can never beat human minds....
I don't have any fear of my job.

Martijn's picture
Martijn
1896 pencils

So one of the best creative directors in France and 6 other people worked on a project that took 9 months to prove they can make better advertising then a machine made by... themselves! How stupid is that?
They made this thing to prove a machine can't be creative, that's like trowing a stone out of the window to prove an object can't fly!

This is probably the most stupid experience I've ever seen in advertising and I really think their clients should ask themselves what they are paying for.

ivan's picture
ivan

I don't agree. Such experimentation is great and you never know what comes out of it.

Martijn's picture
Martijn
1896 pencils

I totally agree an experiment like this can be very interesting, but not when it's made to prove it can't succeed.

Guest's picture
Guest

I think this can be used as a wake up call. What does it say when a robot can perform a job that is so reliant on human interaction? As a journalism student focusing on advertising, reading this article showed me how important the steps are beyond developing the message. I don't know about you, but I want my work to go above and beyond something generated by a computer that only involved a few simple clicks.

Thanks for the article, great information.

byzade's picture
byzade
131 pencils

agreed

agatur's picture
agatur
80 pencils

i don't agree

ehoneybees's picture
ehoneybees
4 pencils

I do not think it's a good idea

ehoneybees's picture
ehoneybees
4 pencils

I do not think it's a good idea

Log in or register to post comments