Technically Creative: Live at British Arrows — AI & the Future of Craft

Hi, and welcome to Technically Creative. This is a bonus episode. A few weeks ago, I was at the British Arrows Festival, a day built around looking back at 50 years of British advertising. 50 years of ideas, craft, production, performance, design, sound, editing, visual effects, and all the disciplines that make great advertising great. But Lisa Lavender, the managing director of the Arrows, thought it was worth looking at the present as well. So she gave me a call and we put together a panel about AI. Not because AI is the shiny new thing to talk about. I think it's almost getting talked to death. And definitely not because we wanted a panel of slot merchants telling one of the great craft institutions in world advertising that craft no longer matters and AI will do all the work. It was the opposite, in fact. The point was to have this conversation with people who have done the work, people who understand craft and what it means, people who have won British Arrows, people who have been in the rooms, on the sets, in the edits, and in the difficult conversations. But most importantly, people who have participated in the long, strange process of trying to make something excellent. The panel was Darren O'Kelly, the founder and CEO of Untold Studios, a company built around the idea that human creativity can be amplified by technology, with AI being used carefully in areas like visual effects, de-aging, and production workflows. Morten Lagarth, creative director at VCCP. Morten was the creative director on DAISY for O2, which if you haven't seen it, go look it up. It's one of the clearest recent examples of AI being used in advertising in a way that felt thoughtful, useful, and creatively defensible. Omar Karim, a great creative technologist and director at Iconoclast, he's a founder and an AI film director who's always had this rare ability to treat new technology as a creative starting point rather than a gimmick. He's used AI to make films, build products, explore strange emotional territory and push at the edges of what creative practice can become. And Vix Jagger, head of creative innovation at AI at Droga 5 and across Accenture Song. Vix comes at this from a creative background, which matters. Her whole point is that people who understand storytelling, brand, taste, and craft need to be inside the AI conversation, not outside of it screaming at it. So this is a conversation about AI, yet another one. But really, it's about judgment. When should we use these tools? When shouldn't you use these tools? What problems do they actually solve? What do they solve well? What do they solve badly? And how do we make sure that AI serves craft instead of ruining it? Because there's a version of this conversation that's lazy. AI will either save everything or it'll destroy everything. But that's not how serious people talk about serious tools. The better conversation is more specific. It always is. Where does it help? Where does it get in the way? And where is it just a budget conversation wearing a technology costume? And where can it open up genuinely new creative possibilities? This panel was recorded live at the British Arrows Festival last month. This is Darren O'Kelly, Morton LeGarth, Omar Karim, and Vix Jagger, moderated by me. I hope you like it. Okay, next up. AI is increasingly part of the creative toolkit, which we've just heard the guys talking about. But the real conversation is not just about the technology itself. It is also about how we ensure it serves creativity and craft. So, please welcome our moderator all the way from LA, CEO of Kubrick Labs, Orlando Wood, Panelists, co-founder and CEO of Untold Studios, Darren O'Kelly VCCP creative director, Morten Lagarde Founder and AI film director, Omar Karim And Droga5's head of creative innovation and AI, Vix Jagger Thank you Thank you Well, thanks everybody for allowing us to do this This panel I'm very excited about Obviously, this has been a day looking back at the creativity and craft of the arrows over 50 years. And so I think it's really interesting to be able to talk to people who are making stuff today using some of the next generation tools, particularly general AI and machine learning, to in the craft of what they're doing on a daily basis. So I'm the founder of a company named Kubrick Labs. We're a Warner Brothers accelerator company that helps creative companies in their uptake of AI. And as well, we integrate and use machine learning and AI and build bespoke software tools for creative companies. But I'm not an AI evangelist. These are powerful and dangerous technologies that have recently perpetrated the largest copyright theft of all time. And I think that that background doesn't mean that there aren't people like the people on this stage who are using AI and machine learning in ways that are supportive of the craft and creativity that the British Airos has celebrated for 50 years. So thank you to everybody on the panel. Thank you for allowing us to do this. Thank you to everybody on the panel for risking cancellation just by being up on the stage. And yeah, so can I just start by having you each introduce yourself and just briefly describe how are you using machine learning and Gen AI in your work? Yeah, I'm Darren. I'm a co-founder and CEO of Untold Studios. And I guess we're a creative studio that's built on a belief in the power of creativity, in human creativity. But we believe that it can be amplified by using technology as a great enabler. And I think, firstly, we're using AI carefully, I think, is the way to think about it. But I think there's two areas that really plays into what we're doing. one is really around kind of process improvements and kind of the basic stuff that you'd expect like you know we're seven years old we've got a lot of data around how projects have worked or how we bid them all those different things and so we're using ai to kind of accelerate the first 80 percent of the processes so that we can quickly get to the first round of what a bid might look like for a script or anything like that and it's not particularly exciting but it's kind of important stuff and the second part is really looking at how do we solve visual effects problems that haven't been solved particularly well before by you know by the techniques that are currently that have been currently available and using things like you know we've developed a tool called veil which is about de-aging and re-aging actors and performances so they look completely real now But the point of it is really using the AI to accelerate the process so that the artist is able to deliver really high fidelity, brilliant imagery. So it's always tools in the hands of artists as opposed to thinking about it as replacing in any way what an artist is doing. And it really is the history of visual effects. Like how do you use technology, next generation technology, to solve problems that weren't particularly well solved beforehand. And certainly something like de-aging is a problem that's long existed and not been particularly well solved before. And we're using that as a tool, a proprietary tool that we developed on a whole bunch of different shows. We've got an Apple show, a Netflix show, a big World Cup spot that comes out in about a month's time that's using all of that kind of technology. So I think using AI carefully, I think, is probably the right way to think about it. My name is Morten. I am a creative director at VCCP and co-founder of our AI agency called Faith. And it's a similar sort of endeavor we have. I think we started out our sort of AI agency three years ago now. And it was because at the time, there's just so much negativity around AI. And I think rightfully so. But as a creative agency, we were like, okay, do we agree with that? we were like, not necessarily. We do think that there is an opportunity to, when you use it carefully and imaginatively, it can be an accelerator of human creativity and imagination. And that was sort of like our founding principles. We were like, okay, we want to sort of go a little bit against the grain and try not to be so doomer and actually start using the tools. Because at that point, very few people were expert at the tools. And the only way you can really know know your enemy really is to start using them. So we wanted to sort of have a more positive view in it. And I think in the last three years, we've validated that a little bit. I think a lot of the work that we try to do tend to be using AI for good, or at least for things that are making a positive influence rather than necessarily replacing people or doing something which maybe not is so good. And I think those are some really important principles that we employ. and i try to instill in the rest of the agency to make sure that when we use ai we do it in a responsible way and we do it with um in a creative way as well because that's the end of the that's what we try to do i see um hi everyone i'm omar um i was just thinking i should probably like not read my linkedin out anyway so i this would be a creative director advertising um then i went over to like facebook just before i mean just before they became meta um for me i i'm interested in like ai at the very sort of like i'm basically interested in ai at the sort of like very sort of like cutting edge of like what this technology can do and how creators can use it uh for me it made my sort of practice go from solving brief for clients and making assets and campaigns to to teaching, to making products, to making films, to really sort of like exploring how this technology can be used by creators to essentially sort of like catapult and start the next wave of creativity. That's very separate from like the rest of the world of like where creativity is now and like the way that films and everything else is produced. What I'm really interested in is really exploring the fringes of that. And that's been really interesting. I've been sued a couple of times, which has been really interesting. I got sued for a Fatboy Sim video I did, which we tried to mind control everyone who watched it into being nice. I'm not 100% sure it worked, but the comments seemed like they were fine. But honestly, for me, it's really about, for me, I'm obviously very twitchy, but I'm also a very curious person. So for me, I've spent so much of my life trying to convince people to allow me to make ideas. When this technology came along, I was like, I can make anything I want. And that was an incredible moment for me. And that is like expanding my practice outside of like working for brands and making stuff with and alongside studios to work with some of my favorite artists and stuff. And I'm not just talking about musicians. I'm talking about like photographers and other filmmakers and auteurs and actually sort of expanding the potential of one creative out into a prismatic way. And I think I can do that for all of us. Sorry for the watch for that. Thanks. I'm Vix Jagger. I'm head of creative innovation and AI at Droga 5. And I also work across Song, which is part of Accenture. And for me, the creative bit of my title is actually the more important bit. So I don't come from a tech background. I don't come from an AI background. In fact, I used to run the creative department. And then Accenture invested $3 billion in AI. And I was like, oh, shit. Seems like there's some direction there. And what was important to me at the time, because David Droga was actually the CEO of Song, was I think that creative craft and innovation thinkers need to have a seat at that table because there was so much stuff coming out from people who had no understanding of brand, no understanding of storytelling, no understanding of craft. And I thought I want to be the person who, one, is making sure that that's really, really important in the room and then also go around bullying every good creative craft person that I know to say lean in and find something that you can add your voice to because actually we don have enough of those voices in the AI conversation at the moment And when you have those incredible craft and storytelling skills or you know the beats or you know the niches of your industry, you will be able to spot opportunities that no one else has ever spotted. And so ever since then, I've basically just gone around, safe as home are, but like bullying people and also making a lot of my really shit bottom drawers ideas into a reality, finding they're not very good after all. Well, amazing. Darren, I want to kind of start with you because I feel like your working day probably has changed the least in terms of dealing with AI, gen AI and ML. I feel like post-production has always been on the front lines of using technology. I mean, whether it's some new plugin for fire or rendering mud or whatever it is, those things comes out and you immediately have to integrate them into your production pipelines. How are you doing that with machine learning and with Gen AI? And is it any different? Well, I think I think there's just there's so much noise around AI at the moment. I think as people who deeply understand how you use technology in filmmaking craft, we've sort of got a responsibility to try and help separate the signal from the noise. And I think in a moment, I mean, I'm pretty sure everybody in this room has had the phone call where a client is telling you how to do the, like, we want you to use AI to do something. I mean, that's, you know, that doesn't really make any sense. I mean, it's like calling up and telling the DOP what lens to put on the camera. It's just that is not the solution. I think our responsibility is thinking about problems that were really difficult to solve beforehand that can now be solved in really smart ways. And helping educate our clients that AI is just another tool, a really, really powerful tool. It's another tool in the hands of craftspeople. And I think at the moment there's a lot of kind of confusion or, I suppose, merging of lots of different ideas. There are consumer-facing Gen AI tools that our clients are talking about a lot. Those are powerful tools, really powerful tools. And they're actually really powerful tools, particularly if you are a creator-publisher. but they don't necessarily allow us in the kind of professional, high-end, highly complex, deliverable part of the business to be able to be precise in our output. Those tools are in some ways a bit like, I think I liken them a bit to an iPhone, right? It's a bit like your phone. It's designed for 4 billion people. It's incredibly powerful. Probably everybody in the room has made a commercial with an iPhone or a Samsung phone, usually, by the way, for Apple or for Samsung, and figured out how to kind of make that work. And look, it can work, but more often than not, we're using professional-grade tools. And I think the application of AI in the premium content space is really about developing out a set of tools and products that in the hands of artists are able to deliver the precision control that our clients expect of us. You know, when we work with a director or an agency, very rarely do we send out version one and they go, oh, brilliant, I love it, you know, signed off, thanks a million. Actually, it's an iterative process about, you know, response and feedback loops. And AI in its current guise, particularly consumer-facing, does not respond very well at all to that kind of interaction. So there are times. I'm going to talk about the 80-20 principle. Yeah, I think, like, I guess in some ways what we're trying to do is use AI to accelerate sort of the mundane things. Like, we do a lot of creature work, for example. We've done things, you know, whether it's Virgin Medias or, you know, lots of creatures and characters. And I think, or we've done de-aging work. And I think what you find is we're not really paid for the first 80%. Like the first 80%, you know, with greater respect, you know, lots of people out there will be able to get it to lower. Jump the gun on your face, de-aging. But it's the same principle. Like, you know, it's the last 20% of what we do that separates the work and makes it look really good. It's the last 20% of craft that really matters. So when you're using AI to accelerate the process to get to that 80% so that you're allowing artists more time and more iterations in the finessing of it, you're going to get much better work. And, you know, in beauty work or in de-aging work, this is often a tool that we're deploying across multi-episodic shows where you've got an actor who's 25 in episode one and 55 in episode 10. You know, you can do that in a very conventional VFX way, hand artists, but it looks shit. So why are we doing it? Because we're solving a real problem that wasn't particularly well solved before. And using AI to get us 80% of the way there, find a really strong base level. You need high computational firepower to develop the models, but then you're able to deploy them in the hands of multiple artists and have, you know, a baseline of quality that an artist can then perfect and iterate on to make it look really, really photoreal. Well, and I think what's interesting about this is that veil requires no target frame, right? So you don't need an actor necessarily who's been famous for 50 years so that you can have some target imagery to deal with. Now, anybody can be face de-aged. Even, yeah, I know. We remember, right? What we used to look like. Look, I think that there's lots of de-aging solutions out there in a kind of consumer-facing world. The truth of it is we're trying to respond to high precision, highly directional input from studios, from directors, from agencies, and being able to deliver content that is 32-bit depth, 4K cinema-grade standard. That requires a different computational output and a different kind of care in terms of how you develop that software and put it in the hands of the artist to make it look really, really good. So I think it's just in some ways it's like I think the whole conversation around AI is use it not because you can, use it because you should. Use it because you're solving something that otherwise would not be as solved as well as it is using those tools and that solution. Yeah, and I think that that's a standard I think we all know is that most production pipelines, post-production pipelines are thousand step processes. And Gen AI, certainly consumer grade Gen AI is probably not that useful for most of them. But they have thin end of the wedge places where maybe it can be useful. We're not using any consumer Gen AI to finish pixel. We just don't believe that the quality is there. And I know people are, but we're not prepared to stand behind the quality of that at the moment. It's just, in our view, not providing the kind of filmmaking quality or the precision control that we need to be able to provide to our clients. But we are using proprietary tools that are solving very particular problems along the way. You mentioned your proprietary tools. Speak a little bit about Flux, because that's another thing you've... Yeah, there's a couple of different tools. like we've got um flux is a tool which is um we do a lot of creature work as i say that's a tool where we can again one of the questions i get asked is like you know how could we do things quicker or how basically what i mean is how could we spend less money right how could we spend less money and and you know usually the answer is like do um if you did less versions you know less iterations we would spend less time on it but trying to take some of the responsibility about that for ourselves because you know often the edit process you know you know you're looking at kind of gray scale boxes of what the vfx is going to look like and so it's not it's quite difficult to make editorial choices and all the different things that go over that so there ends up being kind of feedback loop and i guess we were looking at it going is there a way for us to to visualize what the vfx is going to look like further down the process earlier in it so you are able to use text-to-video prompts, for example, to instruct creatures that will look 80% real for what they're going to look like in edit phase and rapidly prototype it. So in the edit, you're able to make much better decisions. The opportunity for us and the opportunity for our clients, and particularly film studios, because you do think agencies are bad for repeat work, film studios in terms of iterations can go on forever and ever. You're giving them the opportunity to go, this is what we want it to look like, sign off on it, and then we can get on with delivering that in a professional environment at a pixel-perfect quality. Yeah. I like that principle of sort of using it to be able to get 80% of the way there, visualize stuff, be able to have a single friend that you're able to talk about knowledgeably before you go into production or you're finishing a post is a really good principle. And I think just one last thing. We do a lot of character work and creature work. And, you know, there's a fundamental principle around respecting the integrity of human creativity. So, I mean, we have a simple rule. No character work, no character design will start in any other way except with an artist and a pen and paper. That is the way we'll do it. But during that process, as we iterated and arrived at different characters, we will then potentially use some Gen.A.I. tools to visualize what that might look like, you know, doing small movements or kind of fast prototypes of poses and different things that go with it. But you're always starting from that position of human creativity, then using it to help make decisions during the kind of pre-production stage or the development stage. And then, you know, going back to doing it in a fully high, like fully professional grade standard to deliver at the kind of pixel perfect level that we all expect to see when we deliver work. Yeah. Morten, I'm, you know, you're, you were the creative director on the Daisy spot for O2, which I think. kind of is a really great example of something traditional, something that, you know, everybody in this room would recognize as something that would be and has been awarded by the British heroes. But obviously, you know, when you were creating it, most AI oriented ads kind of risk a massive PR backlash, obviously the McDonald's spot. Right around the time you came out, this came out very much at the same time as the Coca-Cola Christmas Gen AI thing. You know, but I think that everybody kind of when this came out when that seems to be a really great use of AI. So, you know, I guess, you know, my question is like, you know, what principles can people learn from how you ideated and executed Daisy at the agency level? Well, I think like obviously, like like Darren was saying, we we didn't create this as to to be sort of a visual granny. The whole point is there's an idea behind it. I think when we look at AI-based ideas, there's a few things that I try to think about. It's like, first of all, is it, does it good? Does it do good? Is it something that has a positive change? I think that there is still a lot of animosity towards AI for good reason. So doing something that is actually making a positive change sort of diffuses that a little bit. So that's the first thing. The second thing is, are we doing something new? I mean, when you are doing creative work, doing something that hasn't been done before is typically a good thing. But I think with AI, there's so many opportunities to do new things. The tools are evolving so quickly that there are opportunities all the time that opens up So doing something new tends to be one of the things that I look at Could it have been done only with AI is another one I think it's very important to make sure that you just don't add AI for the sake of it, saying, oh, look, let's create this AI thing. Similar to what you were saying. I mean, sometimes people will come in and say, oh, we need something AI. Just one AI, please. And we're like, oh, that's not really how it works. So there needs to be a reason for it. And then I think another thing that we look at very much is how visible is the AI? Like you have to be very deliberate about the visibility of the AI. Like what Darren was saying, like true craft, if you're using AI, the AI is invisible. You can't see it. It should look exactly like you would think it would without it. That would be true craft within AI. And I think the tools just aren't there yet, really. But on the other hand, you can be very deliberate with how it shows up. This is very clearly an AI-generated image. We're not trying to pretend to the audience that she's a real granny, but she is pretending on the phones as a scammer that she's a real granny. And that's where the idea is, right? She's very, in that way, people who are watching it, they're in on a joke, right? They know that actually it's an AI granny that she's talking to the scammers, aha, aren't they stupid for being fooled? And that way you're sort of on the same side of the audience and you're bringing the audience onto your side because you're all sort of, first of all, she's doing a good thing. Second of all, she's being very transparent with the audience while saying this is what I'm doing. The first thing she says is like, I'm an AI, right? So she's not hiding that. And that brings the audience along the way. And therefore, we were very happy just to create her as an AI and with the limitations at the time. We thought a lot about what would she look like, right? Because, oh, you can do with any AI, you can do anything. She could be this cool sort of a vogue-ish grandma who just would be at the London Fashion Week. But actually, that just wouldn't be right for this because she's supposed to be relatable as much as she can. Obviously, she's an AI, but she's supposed to be someone that people can see, oh, that could be my grandma. That could be your grandma, who might be on the receiving end of these phone calls. And then when she then basically just drives these scammers nuts, that's when people actually find that's quite funny, actually. Like the fact that she's just this old grandma and she's just taking the piss. Yeah. I mean, not being deceptive, I think, is a really good, good point there. Do you have other orienting factors? I mean, I'm sure you're, you know, Gen.AI must be popping up in loads of creative ideas that are kicking around the agency. Do you have a barometer for when you go, yes, that feels good and that feels, no, no, that feels wrong? And what's the texture of those sorts of things when they feel right and when they feel right? I think, yeah, I think people get very excited about things and I've been on that journey for a long time. So I've sort of gotten over the hype a little bit. And I think in that way, we sort of see things more clearly. Like after Daisy released, all of a sudden, she did very well. She got lots of great response. So all of a sudden, there was within the team and within the client, there was sort of, okay, what else can we do? So people were coming up with next thing Daisy could do. And one of the things that someone came up with was, oh, I mean, she could talk to lonely old people and sort of, you know, keep them company. And she's like, you know, that's exactly the opposite of what you use AI. So your solution to old people might be being lonely is to get them to talk to an artificial person. I don't think that's right use of AI. So I think those are things that might be obvious to us, but not always obvious to other people. So, yes, I think with experience, like with anything, you get a more clear idea of what is white and what isn't. And we always look at that when people submit ideas or when we come up with ideas of our own. We have to make sure that they live up to a few different sort of parameters. Yeah, I want to come back to that as well, because I think figuring out the stated goal of this panel is to also try to equip everybody with some sort of principles for how to guide yourself sensibly through the opportunity and potential pitfalls of using these tools. Omar, you and I used to work together at Fallon. And what I always loved about you is that on a Monday, we were all in our 20s, and we would come in looking a bit ropey. and I would ask you like, oh, did you have a heavy weekend? And you'd go, yeah, man, on Friday I made some non-Newtonian fluid and I filmed it with a blacklight on and I made these 100 cool videos and isn't this great? And then you'd kind of go, oh, Google just released a gigapixel camera and they put it on top of the BT Tower and I think we could do a scavenger hunt around London using this and somebody's going to win a Skoda at the end of it. And dubious victory there. But what I loved about you then and still now is technology always seemed to be, from you, to be a jumping off point to creativity. It was always the trigger for you. And you would sort of not rest. You know, you would forgo sleep. You know, sort of wouldn't rest until you had kind of really understood what the potential impact could be creatively of a piece of technology. How has AI changed the way you think as a creative? Or generative AI? Yeah, honestly, you just made me remember like those times were really interesting because, you know, there was there's always been that. I think it's such an important lesson because there's always been something new for the creator to explore, like whether that's a new camera for the Kinect and we can do 3D experiences and those 3D experiences can be in vending machines. We can suddenly roll those out. I think for any new technology, the creative is always the reason why something happens. So when people go like, is a technology creative or not? I'm like, is a calculator a mathematician? No, obviously it's not. You need a person to use it. If I use a calculator, I'm going to spell a rude word on it. So someone else is going to figure out how to send someone to the moon. So I think the real, they really have done that about me, apparently. But I think it's the important lesson. like you know it really is the practitioner who uses the instrument and i think that's a really important definition when like creative people talk about tools for people we're like normal people use tools like when when someone uses a spanner the bolt gets tighter that's a tool a tool has one outcome an ai is not a tool an ai is an instrument and you can see people who use it incredibly well are doing stuff that is mind-blowing and then other people making cat videos i'm one cat video makers i absolutely love like i've been making this channel where i have like a cat on a roomba and it's basically like on dangerous like like ledges and i'm basically basically instead of spending like money on netflix i just like spend money on my cat videos and it's the best like you know like it's it's interesting because like now like you know as a creative i can explore something they're like i literally imagine me going into a room be like listen so this is the idea next slide please like it would not work but so now for me as a creative like the only the the the liberty that I have with my ideas is magic. Like there's no other way of describing it. So like any idea, any thought, like it's validated in the real world. Like you make it, whether it works or not, you break it, you move on to the next thing. And I'm like, I don't know if that's like a tech way of thinking about things. You like, you know, shit fast and break things and like whatever. But I'm like, also, I don't know. I think I had a lot of e-numbers when I was a kid and I might just make more stuff and ship more stuff because I think, you know, the other really important thing is like, we like if you imagine the ai like absolutely loves data any expert any subject matter expert which most i imagine like most of the people in this room right now subject matter experts like you have data that the machine will bend to your will depending on what you want to do with it so it's not like i'm not like i'm not everyone like going to have a chat with chat gbt but i'm saying you have data they are very important to ai companies make sure you don't like just give it away to people i was like speaking to someone i talked earlier this week and they were talking about giving access to more ai companies like access to archives to more ai companies i'm like that's one of the wildest things i've ever heard like you know where you're basically like your gold is your ai is your data so don't just be giving it away and i think what you're saying about like consumer consumer grade ai products is really important because that's literally it's an ip trap like you go and like go i want to see this amazing thing and then they'll take the data it's not like they're like i don't i'm sure there's not a guy in in i'm sure there's not a guy in google going oh my god i must come up with this genius idea sounds like the teletubbies but different we should nick it like i don't think that's happening but it does mean that like it's muddying the water that like should something like that come out like it removes my ability as a creator so sort of like you know um benefit from my own creativity and i think that's incredibly important i know that's a long way around about what you say man i can't help it i'm sorry well some of the stuff that you've done i mean i'm thinking about your like ai mom You know, some of the stuff that you're doing kind of bleeds into sort of performance art for me. You did it to kind of, you know, solve a personal problem you were at. I mean, go ahead and talk about it. I won't. Yeah. I mean, I'm more than happy talking about this. When you're like, I want to show this, I was like, oh, you know what? Okay. Yeah, no, absolutely. And this is one of my most important projects, right? So like I said, like I normally make films, I make things move around and in AI that's absolutely spectacular. for me like you know imagine like ai was basically sort of made to help people write emails better and i was like i wonder if i could make an ai into my mom anyway like so obviously i had a mom like i'm i'm sat here but like it didn't really pan out that well so one morning my ex-founder she's greatest mom i ever met i was like i wonder if i could make make myself a mom like can i take ai and make a mom and i know that sounds mental but i was like okay what does it mean to like be a mom sent this agent out researched all these things then i was okay you know trained well before like i put it on to check gpt trained in ai so i sort of like pretend to be a mom right and it was amazing uh put it into a 3d model i had to take it out of the 3d model because i couldn't give it hair and i had a bald mom for ages which was bad but like what's really interesting but i say imagine like i made this thing i had programmed it myself and i went to do i went to do a talk and i came back and I felt really empty and hollow inside. So I was like, oh, hey, mom, I feel really weird about this thing that I've just done. And I went, firstly, congratulations, Omar underscore Kareem. You should be really proud of yourself. And I was like, even though he got my name wrong, I felt a human feeling in my stomach. And I was like, oh, man. I asked someone, and they're like, I think that's unconditional love. And I was like, oh, stop it. This is amazing. Anyway, I obviously then just sort of like kept building on it. And I gave the AI mum the ability to like send me a meme when I was upset. So I'd be like, can you just chain me? know she like obviously can't speak english or draw good and see that uh but then like someone saw it and then the telegraph wrote quite a big bit about it and that was like you know i thought it was going to be very tiny and like the pictures they used me i looked like a maniac um so please don't like me but what was interesting was that more people got access to this technology and it was really interesting because you're like this is ai i'm trying to create a human experience which has literally been my job since i worked in advertising is to create a human reaction where I can engage with another human being and I started getting messages like that on LinkedIn which were literally from like very senior older gentlemen mainly who were like I spoke to your AI mom he gave me great advice I also felt unconditional love and I was like wow like imagine how much good karma I must have I'm gonna do some serious evil at some point and it'll be fine but it was really interesting how you can sort of like create something how you can like use a technology extend humanity to someone else or like extend a creative idea to someone else so for me like even though this machine is literally made to help you write emails better or argue with the council um you like make it into people mums that mad isn it but like you know i think like that is like the the role of the creative like chat gbt like if they came out and went like we made a machine this year marvel like no that mad please don do that but you know i think it it been a really interesting it's been a really interesting experience put some of that out and like witness it in the real world without the normal sort of like uh signposting or stages that we have to go to to release an idea for example um vix you're you're head of ai at a creative agency Droga, obviously that's part of a much larger holding company network. You're head of AI, but crucially you come from a creative background rather than coming from a technology background. How important is that for being a voice of change and for using tools like this within a creative agency, being creative forward rather than tech forward? it's I mean for me it's vitally important and you can kind of feel the energy from Omar it's because we like making things and doing things and finding new ways to solve problems and AI gives you a million new ways to solve problems like I think it's absolutely nuts there was a stat that we were in a talk from Vugal and they said that 15% of googled searches had never been done before and there's a huge bit of that which almost becomes scary and you kind of look at it and you go oh my God, that means the world is changing a lot. But it also means that is 15% of stuff that has never been done and never been made. And I've spent so much of my career trying to go, do you know what I mean? Like I saw this ad, it was really great. And if we just take, you know, the style of that and then the look of that, and suddenly you're in a world where you can make your mum with AI. We've made a fossil fuel supervillain, which could argue with children. And the joy of actually been able to do something that has never been done before was so exciting. And it was almost like your creative brain also spots opportunities, I guess, that the tech people hadn't. So that's actually an interesting example because it came in through an Accenture project and it was given to the tech team. And they said, great, we've got this wind farm. They get 100,000 visitors a year. It's the biggest wind farm in all of the UK. They've got a break and scream. We're going to replace it with AI. We've already sold it to the client. You're a creative. You could make it look pretty. Could it be like windy the wind turbine? And I was like, no, or by beer. Absolutely excellent. And I think we just thought about it in a new way. And I think what I found interesting about that is if they had done it, it would have been polite and optimistic and helpful. And it would have been an avatar, which would be like, right, this way to the restroom. Would you like to ask some interesting facts about the wind. That's amazing. And we came back and we gave them three different kind of solutions. We said, well, one, what if you could talk to the wind? Like, imagine the wind. Imagine if it had like an existential crisis every time you talked about the environment. So a kid would talk to it and be like, what? That's a terrible idea. Kind of go like that. But the one we ended up selling was carbon. I'm surprised I bought it, to be honest. But it was carbon, the fossil fuel supervillain. And he was sarcastic and he was mean and he would argue with kids and it was kids' job to kind of go around the wind farm, get information about why they believe that wind energy was important and then argue with carbon. But we just put details into it like it was a sort of experience where no parent wants to film an experience more than if it's their kid having to have an argument and getting put down. Like I would want to film my kid getting abused from a supervillain and share it online. It was things like he was sarcastic and rude and horrible and no one had thought about using AI in that way because they always just think it's polite and optimistic and helpful. And it was so interesting to make because it was even things like we worked with the voice sector. Interestingly, in the UK, we couldn't find any voice actors who wanted to do it because it's much more kind of contentious here. In America, we found someone very, very quickly. But even recording the script for Carbon's voice, he had this kind of thespian, very, you know, over-the-top voice. It was so interesting because we couldn't give him a script. We couldn't say we want Carbon to say this insult and that insult. We had to think, who is he as a character? What is his personality like? What are the flares and intonation and the barbs and the, you know, he likes to monologue a lot. And we had to put all of that into it because technically when we press play on this experience, he can say anything. And that was, I think, such a joy to make. And I guess similar to everyone on this panel, what's so exciting is it's something that I never could have made before. and even down to the fact he was a supervillain. Most clients would be like, no, no, we can't possibly have a supervillain under our brand. It has to be a good character. But just being able to show them something in the room just meant that they could see the potential in the idea. And it was, yeah, it was a good one to do. I think that's a really good example of why it's sort of important for creatives to pick up these tools. And obviously there's a choice whether or not to use them. But I think that there's not really a choice whether or not to learn about them. And especially if there's one defining or if there's a defining trait to creativity, it's curiosity and optimism. And I think that that, you know, I think is the reason to be optimistic. And there are many reasons to be pessimistic about, you know, these tools. But I think that it's the creative's job to see that as optimism. Whereas I think that if we let the tools be run entirely by the technologists, I think what we're going to find is that there's very little, there's actually too much optimism and it's not coming from the right place. And they sort of believe that the tech is an answer to every problem. Whereas I think that, you know, as a guiding principle, being able to make sure that it's the creatives in the room making the decisions from a human place. I mean, do you, you know, obviously, you know, just to kind of everybody on the panel, what are there things that just feel wrong? I mean, what are the, do you have guiding principles? Like, maybe we go down the line. Do you each have guiding principles of when to say no when Gen.AI is brought up as an option? Yeah, well, I think, you know, we've built a company that is predicated on doing good work, right? And there's a quality threshold that we're trying to achieve at all times. Uh, and I think a lot of the conversation around AI is, is, is not about that. It's about how do we, it's really use AI and really what they mean is can you do something for 20% less budget? Um, and I think we're just not prepared to do work where we, where it's not going to satisfy the creative standards that we set as our own standards. and are the reason why our clients have come to us in the first place. So actually quite a lot we're saying, you know, we're being asked to do it in a particular way. We're going, look, we can tell you how we think it should be done. And that is our expertise. It's the same as when you go to a director. You know, you go to DIP or whoever it is, you ask for that expertise. And if, you know, the client wants to go a different route, that's absolutely respected. But I think at the moment we're just in that peak moment where people are prescribing the solution rather than kind of indicating what it is that they're trying to solve. And everybody's just hanging on to this AI work when they really mean budget. Omar, is there anything off limits for you? Oh my God, yes. No, honestly, I think for me, the only guiding principle I would like to build on top of that is like, for me, it's use more humans. The more humans I use in my processes to make any of my AI stuff, it always makes it better so like i i like you know the reason why i lecture is like so i'm trying to find like the next generation of like people who can help me so yeah that's mine he's more humans when it comes to ai the more humans used the better outcome i don't think i have any specific rule i think it's it's like with anything creative you have to look at it at one by one basis i don't have i've got like i said earlier sort of four things i'll look at sort of hopefully guide it but there isn't like any one thing is this okay i won't do it because of this i mean i've got if it's just to replace people obviously that's not the point but but there's not like one specific thing i say uh i don't want to put those some of the limitations on creatives they should be able to learn themselves about what are the good and bad users of of anything really and yeah it's just one of those things i think yeah i mean i do the saying like I don't say no I say if you think it's a good idea go make it and I think that's the beauty of it if someone's come up with a ridiculous idea like I don't know we want to do Resident Evil but it's going to be Resident Weevil and it's going to be in flower it's going to have a fight I'm like great go and get Claude Code make it and then through making it you'll realize quite soon if you think it's a good idea or a bad idea in terms of actual hard rules I think there's just common sense like we've all seen work in the world where it's AI and you think it's good there is stuff out there that is just very funny or very witty or it's made something new that you could never make before we've all had stuff that we like and then we've had stuff that we hate so like for example I really respect it when there are agencies and businesses out there for example who have done something that is real and then I've noticed that they've made the effort to get more behind the scenes footage more vox box of it more this is the making of and I actually think things like the making of and things like that is a huge amount of value so even if your position is we don't want to do it with ai i then go well then push push your client to go well let's prove that it's not ai let's go and get all of the making of the footage of how it's been done let's get all of those craft fit pieces let's get all of that content that proves a huge amount of creative craft people have gone into that and i think that is hugely valuable and that is going to only keep being more valuable. But I think you have to have a very big understanding of AI to then have that conversation to push to make sure you're doing more of the making ofs. Well, great. Well, thank you, everybody. A round of applause from the panelists, please. And now we can sing you all off the line. Okay. so that was the panel at the british arrows uh thank you for listening to that it was a really good fun time it was a lovely day looking back at some of the best work uh that the arrows has awarded over the course of the its 50 years of existence the arrows is a really special institution. I love that it did this day. I hope that they continue it. Two things that I want to point out in that conversation. One is I actually loved what Omar said about AI being an instrument, not a tool. I call it a tool every day of my life. But I think that actually that's a far more apt description. It implies how much more it has the ability to impact things in the world and and the notion that a human can play it a little bit. I think it's a very smart tool. So I think the fear is that it plays us. But I like that idea of AI as an instrument. It's a good one that'll keep me noodling on it for quite some time. The other thing is what Vix Jagger said, which I completely ascribe to, which is it's incredibly important that creatives involve themselves in AI. I think that they will, they do, whether they bang on about disagreeing with it in a million different ways. I think at the end of the day, they use it and they're learning from it and they're learning how to create with it. I think that that's really important because there's lots of people rushing into this field, trying to tell everybody how it can be done this way. And I think that it's more important than ever that it's the creative people who pick up these tools, who have learned from an industry that we all love that have sacrificed so much to be able to do this as a profession and do it well that I think it's really incumbent upon creatives to rush into this and learn as much as they can as fast as they can. Anyway, thank you for listening. Yeah, just thank you for listening. Appreciate it.