For today’s Explainable I spoke with Jason Carmel, Global Lead, Creative Data at Wunderman Thompson. The agency’s Subconsious Order campaign won the Creative Commerce Grand Prix at this year’s Cannes Lions. The promotion utilized AI to guide users of a food delivery app into selecting the right order. Jason and I chatted about the campaign, how he sees advertising and AI evolving and why AI tools are like a garlic press, an analogy that I will be stealing. The interview has been edited for brevity.
Explainable: Let’s start with the Subconscious Order campaign. Can you tell me a little bit about the background of it and how you got involved?
Jason Carmel: Hunger Station is a client of ours in Riyadh, Saudi Arabia, and it's a food delivery service. And I would say their biggest competitor isn't another food delivery service, but it's people rage quitting and just deciding to eat what they already have. So they wanted a mechanism to help connect with their audience in a fun way, and reduce the time for decision-making.
So we've been messing around with AI for, I want to say, about six years now, give or take, doing various projects both generative and well beyond generative AI. And we found a technology that actually used computer vision and allowed really accurate eye tracking in a way that we thought we could use both on the desktop and on the phone. And so we had this idea of what if we could let somebody just look at a bunch of pictures and use that to determine really what they were in the mood for. And then when we found out that Hunger Station had that problem, it really seemed like a match made in heaven. So we shared with them the technology, they helped package it up into an idea, and the client fell in love with it.
Explainable: For a lot of people,AI tools kind of became a thing on, like, November 30 last year when ChatGPT landed. As Wunderman Thompson’s Global Creative Data Lead, I assume AI has been something you've been thinking about a lot in recent years
Jason: Yeah, I mean, AI has been around for a long time, and I think it's been accessible in our industry if you looked for it for maybe the last decade or so, but you really had to look for it, whereas now you can't swing a dead cat without hitting generative AI somewhere.
One of the things that I secretly, as a practitioner, love about the Subconscious Order is that it uses two different types of AI that are really hidden from the user. This isn’t ‘an idea about AI’ idea. It is an idea that I think is compelling on its own, that uses artificial intelligence to just make the idea seem more magical. So we use computer vision, which tracks your eyes, and then we use a topic modeling artificial intelligence, which helps us understand. When you stare at a picture of fried chicken, is it because you like fried chicken or you like fried food, or you like chicken, or you like soul food? Like what is it? And then what do we show you next? But all those things are kind of hidden beneath the idea itself, which I think will be more of the rule, rather than the exception, going forward. I think the era of like, ‘hey, here's something we made with generative AI’ is going to be over soon, if it’s not over already. And I say that as a huge fan of generative AI.
Explainable: What kind of AI tools have really helped you in your job over the last number of years?
Jason: Well, I think we use a lot of the broad platforms that are out there. So I mean, this isn’t going to be shocking to anyone, but we use a lot of Google's TensorFlow stuff. I think the Microsoft Azure packages have been really interesting and both of those teams have been very happy to get developers playing with their stuff. And then as you start to get these large language models basically performing the same role you have with OpenAI, trying to get people to experiment with it, with their stuff, I think that's where we focused a lot of our attention from a tools perspective. It's really just pulling the tools that they have available in these major platforms and making use of those. I would say that's like 85% of what we do.
Explainable: And over the last twelve months, have you seen a sudden uptick in people across the agency who are suddenly talking and excited about AI?
Jason: One hundred percent. I think the hype cycle of AI, as much as people kind of roll their eyes at it, is also a very good thing for the agencies at large because it gets people interested in technology. It gets people thinking about what's possible. And I think when you get a creative organization thinking about what's possible, that leads to interesting work. For me, it's been net positive. I mean, obviously, you get the ‘could we just magically do something with AI?’ and it's not possible. That happens for anything, but I'm actually a big fan of the fact that everybody has their eyes on this. I think it's overall a really good thing.
Explainable: It feels like it's the new ‘can you make this go viral?’
Jason: Yeah, it is a little bit. I have a little bit more faith in AI's longevity, only because it's already part of so much technology that we're beginning to use. Our cell phones have it, our computers obviously have it, our cars have it increasingly. And so people are using it whether they believe it or not. As opposed to something like, I don't know, an NFT, which maybe didn't have the staying power that some people thought it might have.
Explainable: It kind of feels like when I speak to people who have kept abreast of AI developments for longer than twelve months, for years and years and years, they have more faith in the longevity of AI, but lower expectations in the short term.
Jason: Yeah, I still view artificial intelligence as a thousand pretty good tools, as opposed to one stupendous tool. I think that's a misperception that a lot of people have, that AI is just one global thing, that there is an AI that does all of these things. As opposed to a million AIs that do very specific things. So I view artificial intelligence as like a garlic press. Which is almost useless to anybody who doesn't use a lot of garlic, but to somebody who uses a lot of garlic, it's awesome. We have a thousand different projects that have a thousand different specific applications where we need a garlic press. And AI has been amazing for that. But I'm also not worried about Skynet taking over, any more than I'm worried about the garlic press taking over.
Explainable: I like that analogy. Do you feel like grasping AI tools is essential to being creative in advertising now? Or will it be a strand of advertising, but not all about advertising?
Jason: I don't think there's going to be, like, an AI wing of advertising. I think a lot of the people who do advertising work will use AI tools and they'll make great strides with it and it'll accelerate their work meaningfully. In the same way that I don't think agencies are focusing on digital stuff anymore. Like, ‘we have a digital wing of our agency’, now that the internet is so ingrained in culture, it's kind of antiquated thinking. Or social as well, yeah you do get social specialists, but I think people think about social in the broader context of advertising as opposed to vice versa. I think for artificial intelligence it'll be the same way. I don't think people will be hiring advertising AI people, they'll be hiring people who can do advertising, who can use tools that they have available, a lot of which will include AI.
Explainable: If you were speaking to a creative trying to get to grips with this new world, where would you advise they start?
Jason: Oh, goodness, it depends so much on what they're already interested in, because I can tell them right now, there's something cool the AI is doing in that space right now. And so the best thing that they can do is to just read about it and try and become a part of it. Every tool that an expert already uses has an AI component, whether you're a video producer, or a musician, or a copywriter. So really it's a question of just doing the search and start going through them. Some of these tools are terrible, and they do a mediocre job of what they claim. And so it is okay for a person who is a very good video editor to look at an AI video editing tool and say, this one's terrible. They don't immediately have to doubt themselves because they don't like that tool.
As somebody who plays with AI a lot, I'm fundamentally optimistic about humanity, and I think that it is going to be the humans who use the AI tools rather than vice versa. So I would just tell them not to be scared of it. They are still the experts. They can see good from bad.
Expert Interview is a regular feature where I talk to creatives about their use of generative AI tools. Their wins, their frustrations, their tips, and their concerns. Want to talk about your own use of generative AI for a future issue? Get in touch contact@explainable.online