How to use AI in your workflows without costing the Earth

Hello, lovely human beings. It's so, so fab to have you here. Thank you all so much for taking the time. If you haven't already, do pop in the chat like those absolute legends have already saying where you're watching from and saying hello to the community. We've got people like Muriel who's in Montreal in Canada, we've got Becky in Exeter, we've got the legend that is Joe Royce saying hello from Amersham, I imagine in a different room to Seb Royce. So hello Joe Royce and hello hello it's a pleasure to have you here. Let's get started with today's session which I'm blooming thrilled to bits that we can actually bring to you because this feels like the type of session which if we were just speaking about social media or just speaking about AI, it wouldn't be as important as this kind of session. So first things to say first, if you want to hang out with these wonderful human beings that you see in the chat feature alongside you just there, then do come and hang out in our community space called HIVE after today's session. It's free, it's an opportunity to learn, we hold regular AMAs with incredible people, so you have HIVE right there for you. Second thing second, let's get going introducing our incredible speaker today. So we have, Seb Royce with us. Now Seb has a job title, it is a creative and communications partner at Eden Lab, which is a green growth consultancy and venture studio. But really the reason why I'm pleased to have Seb here today is that he's one of us. He's part of the community. He is a marketer. He is a marketer with an interest in green stuff, and making sure that what we do is paired alongside the planet and making sure that one doesn't come at the cost of the other. He sat there figuring it out and figuring it out in a scene which is changing quite a lot. And that's the reason I love having Seb here. We're not coming under the pretence of scientists or whatever, we're coming under the pretence of a peer sharing everything they've learned. That feels like the spirit of the Marketing Meetup. Today, we're going be speaking about how to use AI without costing the planet. The reason why we put on this session is it's one that the community asked for. Every time we have a chat about AI in the community, there's usually a comment or two which says, yeah, but what about the planet? Well, that's what this is an opportunity to discuss. Today will function as a talk. So Seb has some slides to share at the beginning and then we'll have time for Q and A at the end. So if you have any questions, any thoughts, then do pop them into the Q and A and we'll come to those at the end of the slides and the presentation element. Today, we have a featured sponsor. That company is Canva and we wouldn't be able to bring these sessions to you if it wasn't for our sponsors. So we are so, so grateful. Canva, I don't think I need to introduce the brand, but what I can introduce is an offer. So Canva have three months of Canva Pro available to the Marketing Meetup community if you use the code TMM3 free, which is really hard to say. So if you want to take advantage of that, do head into that code right there. Also, big thank you to Wistia, Mailchimp, Prismic, Front of Fire, Planable and Cambridge Marketing College. We speak about these brands on rotation as we go. Right, that is my introduction done. I just need to acknowledge these incredible human beings in the chat. So I want to say hi to Celia, Emily, Molly, Rasheen, Natalie, Dylan, Lydia, Caroletta, Sophie, Rose thank you all so much for popping in the chat where you're watching from Seb is on screen he's the man with the thoughts and the best hair in the entire world So Seb, it's, it's over to you mate. Thank you very much. That's great. Hi everyone. I'm hoping you can see this and I'm sure Joe will imagine if he can't. It's really lovely to be with you here and, I before we get into kind of AI and the questions around sustainability, I just wanted to give you a tiny background about how I got here because I was probably in the same position as a lot of you are now a few years ago, but I sort of made a choice maybe to to push to the science slightly. I work on a portfolio of things. I used to work at agencies. I was a creative director. I was a founder of different agencies. But, but I've moved and I now work with places like Eden Lab. And Eden Lab are a group of sustainability experts, ex communications people, business model builders, and economists. And this is this is what we believe really, that the transformational organisations and iconic firms of tomorrow are being born right now, and they'll thrive and create jobs and make money without destroying the world around them. And what what we do is we we help them work out what they could what they could become and how they could get there. And of course, lot of those organisations are going to be underpinned now by, by their AI. I don't need to tell you about how screwed the planet is. We are overshooting all sorts of planetary boundaries. I don't know if you've gendered up recently on where we are with climate change and the environment, but, I don't want go into too much, but I would really, really urge you to, write down this web address and see if you can see a film from the National Emergency Briefing. I don't know if you've about it or heard about it, but a few weeks ago, they, they're a group who wanted to do public information initiative, and they got over one hundred MPs together in Westminster in London, and they got ten climate scientists to brief them on the societal environmental impacts of climate change and if it's not dealt with, kind of immediately. It's a pretty stark briefing, but it was also really compelling to hear it from people who really knew what they were talking about with the latest data and updated information, because there's lots of miss and disinformation online about all this stuff. Really interestingly, the film of the event, which they've cut down, into a kind of manageable hour and a half, ninety minutes, I think, is touring the country right now. So I think it's it's showing in Amishon next week. It's in lots of town halls and community centers. You can even request for a public showing in the area that you're in. So go to that address and you'll see there's a map, an interactive map there. You can see where the film is travelling around the UK and it's a really good way of just getting up to date on what's happening with climate in the environment right now. Donald Trump did not make this picture and I didn't find God, but I just wanted to tell you briefly about how I got to where I got to. As I said, I worked in agencies a long time about seven years ago. I had a bit of a moment, like a road to Damascus moment, hence the weird biblical image, when I think I was probably working on on brands that weren't making me feel particularly good about, the products and services that I was bringing into the world. Also, of just had kids and starting to think about the the environment and the place that they were going to live up in. But once I'd had the kind of moment of am I am I using all my skills in the in the right, way here? I started thinking about, the footprint that I was having on the world and everyone knows about carbon footprint. It's a very well known, term, but I started thinking about what my work footprint might be and, obviously had quite a long career in advertising and marketing and I thought what is my work footprint? I wonder if I could sort of wonder what it would look like and I took into account the businesses that I work for, the brands that I promoted, the campaigns that I produced and where I'd produced them, where I traveled, the technology I'd used to produce them, which of course now really includes AI. And I thought, God, you know what? That is a relatively heavy work print, and how do I feel about that? And at the time I didn't feel great with where it was going. This was coupled obviously with working at the kind of like the cutting edge of digital media, which has now got even more precision. This is, call them weapons of mass consumption. This is all the performance marketing tech we have at our disposal, which is awesome and can help us sell more stuff and it's, it's amazing. But if it's focused in a slightly wrong way, it can just drive rabbit over consumption, which is what is happening, at the moment. So all these things were considered and, I thought we needed to take a different path. I sent myself on the Cambridge Institute for Sustainable Leadership course. If you ever want to do a course in sustainability or sustainability in AI or understand the the influence of the two on each other, then the Cambridge the Cambridge course is really, really good. Loads of people who are in marketing or still are or advertising have done the course. It's really good that you can do a six week one, you can do a three month one or a longer one, it's quite intense. But it gives you some real tools and makes you realize that you could be part of the climate solution, as well as potentially exacerbating the problem. This is what it comes down to really. With our ability, your ability working in marketing or advertising, we, we create good products and we create good services. We can create really compelling narratives, and we can spread ideas. Then we can affect behavior change. And those are really, really powerful things. As soon as you step outside the advertising marketing world, you've realised quite how powerful your skills are in communication and being able to create narratives. But of course, you know, you need to be able to use that power responsibly and focused in the right direction. So that's kind of how I go ahead now and why I think it's really important for for all of us really to try and do what we can, within our organisations to work and live a bit more sustainably. But sustainability isn't having a great time at the moment. This lovely design slide, sorry it's a bit messy here, says a lot of good kind of a lot of things about that's going on at the moment actually. Seventy five percent of UK consumers say that rising living costs, are changing how they shop. That has a massive impact on sustainability. There are quite a few sustainability only brands, so brands that marketed themselves just on sustainability that have really struggled recently. And at the moment, people want price. They want convenience. They want availability, and all those things need to come first. Sustainability needs to be there if we're gonna have better products, but it can't be the only thing, and it certainly can't be at the expense of those other things when money is really tight. The lovely guy in the middle, I don't want get too political, that's Richard Theiss. So you'll be seeing a lot of him in the next few months. I I don't doubt. He coined the the term net stupid zero, and a lot of the sustainability stuff that we're hearing now is being weaponized and politicized. And it gets all mixed up together. You know, there are some policy stuff. There are some broader church, expressions like saying sustainability, and they all get wrapped in together. And that's not necessarily a good thing. You'll know the guy on the right. He doesn't really believe in sustainability. He thinks it's all a bit work. And what happens is that brands and organizations start green hushing, that's the little zip guy in the corner there. They stop talking about their sustainability initiatives. That's really dangerous because people don't think that there's lots of progress going on when there actually is, but also people can avoid, the spotlight being on them. Organizations can avoid being the spotlight on them. So it's a funny time for sustainability. At the same time, you've got AI starting to power and shape and turn the whole industries on turn every industry kind of on its head. There are still a few of these guys roaming around at the top of brands and organizations. I call them CEOSaurus. They have a different attitude to sustainability, kind of prehistoric attitude. They talk a really good game on sustainability, but they're thinking quite short term. They're thinking more quarter to quarter. They can't really see past their AGM and they see sustainability, and lots of people do still, I think, as a regulatory burden and not as a real commercial opportunity. There's a whole another chat we can have about the commercial opportunity of sustainability, but it is real and it is there. And they also think that sustainability is only about what you say, what you claim, not what you do as a company, and that is really important as well. So, they tend to hide out sustainability to the sustainability department, which may be a couple of people definitely not C suite and definitely not part of their core business. So, the reason they're dinosaurs is that they will eventually be extinct. It's just that they're still there and they are blockers to some of the progress that we need to make. Let's be a bit more positive now. At Ethanolair we talk about the regeneration and we talk about it not as a demographic of people, but as a mindset across different industries and across the world of people who want a cleaner, more affordable and a fairer future, and are doing something about it either by what they're doing at work or how they live or preferably both. And we all have a part, we have a chance to be part of the regeneration. It's not about how senior you are and it's not about what industry you work in, but it is a mindset and it's a mindset which is spreading and growing and it's really exciting I think. Your planner does need you, but what can you do sitting in an advertising or marketing agency? Realistically, what can you practically do? I started thinking about conscious creativity, a while back and defined conscious creativity as, an approach to marketing and advertising that is aware of its societal and environmental impact and does what it can to mitigate those worst the worst of those environmental impacts. It has a central premise to, conscious creativity which is starting with the end in mind. This is a really important, principle because it's able to take every single decision that's made from then on. Many sustainability initiatives that seek to minimise environmental and social impact are about offsetting damage, once campaigns or things have been done. But adapting a conscious creativity approach allows you to make better decisions as you go, rather, than try and offset them at the end of things. And one of those will be to do with AI. So, we'll get onto AI properly now. I've got this I'm torn on AI. I love AI. I can make as a creative person, I can make amazing well, I think they're amazing things compared to the things I used to be able to do. Not just images and content but films as well and I'm trying to focus them in in a direction which I think is positive and good, and productivity has obviously gone through the roof. I'm not one of those people who thinks that AI is better, for improving my creativity, but it's improves the amount of things that I can do, and I love it for that. And I know that this is just the start and it's going to be transformative. On the other hand, and I just talked about films and content, I'm slightly daunted about the social and environmental impact. I'm sure some of you are too, but I've got friends who work in production, who work in film production, who work in advertising production, who haven't worked for a long time now, because the work's dried up, because you know I can do the whole lot from from my machine. There's obviously massive social impacts that are going to be happening as well and the pace of change is so fast that I think is a bit kind of daunting, you know, for everyone. Angelina said that, you know, AI is like electricity. Just as electricity transformed every major industry a century ago, AI is now poised to do the same. I think that's inarguable, and I really believe that as well. This is a huge, huge monumental, shift, the surge in energy energy demand driven by by AI. Ironically, electricity becomes the issue as well, as well as water, and other things. So, let's talk about the scale of the problem. Again, you might have read a lot about data centre use of energy, about water usage and how much water data centers take in. The scale of the problem is pretty real. Every prompt and image or analysis that we generate requires energy intensive processing or compute, on remote servers. And of course, as global AI adoption accelerates, so does the strain on power grids and water supplies and the raw materials to build the GPUs, the general processing units, and the hardware kind of behind it all. As I said, there are millions of stats. You can choose your own. These ones kind of let out at me that if global data centers were a country, they'd be the fifth largest energy consumer in the world, kind of somewhere between Japan and Russia. That's now. And, the projections for growth are astronomical off the top of any graph that you can possibly imagine. A single large data center can consume as much electricity as a small city. And of course, they drive even greater demand than conventional cloud computing. And so, if this electricity consumption doubles by two thousand and thirty, you know, where do we end up? By the way, we don't have to have, solutions to this right now, but the the but the facts are pretty stark. And also water, which is a less discussed dimension of data centre resource use, is incredibly important as well. It's used to cool all of the systems of the data centres. So should we worry? By the way, every image in this presentation is a recycled image. They're AI created, but they're from existing presentations. I didn't want to do more prompts and create this, so these are from existing presentations. I have no idea what this is for, but anyway. So, be boring. The answer is really that it depends on scale. So one of us writing a brief on Claude isn't really the issue, it's the cumulative effect. So if you have a team generating ten thousand, twenty thousand AI images or video, a week perhaps regen leaping on really vague prompts. I can't talk about that at the moment, but we're still experimenting with prompts all of us. We're not particularly good at it. We don't do much prep and we get them vague and wrong. Using frontier AI models, large language models for basic tasks, I'll talk about that in a second as well, that becomes a different conversation because the cumulative effect really does build up. If you have a you're working in a big organisation, that can that can increase your carbon propylene so, you know, massively and very quickly as well. So, what do we do? Well, these are my own personal take on things. There's not a scientific reason behind some of this. Actually, that's not true. There is some scientific, reasoning behind it, but I wanted to give something which I think are some which I think would be more practical and easier for you to execute against potentially. One is rightsizing your your AI models. I like this image, this kind of Formula One car. Someone said to me that if you use something like Ford or Jack GPT for real simple tasks, then it's a bit like driving a Formula One car to the shops. You just don't need a model which is that big and hungry. What they can do is absolutely insane, these large language models. What we use them for is a tiny fraction, really, most of us, and what they can do. And there are smaller, lighter models that are available, but they're just not as well known. A lot of them actually are subsets or they're tied to some of the bigger, large language models. It doesn't take much to actually dig in and find out which they might be, and if you're working with an organisation, you might well have a policy which allows you to look at right sizing your model together that it has a cumulative, impact. The second thing is to write better prompts. As I said, we're just experimenting. We're all new to this. We're not very good at it particularly, and every time you hit regenerate on a really vague prompt, you're spending carbon, not just time. And so better prompts are really a sustainability intervention. I think how many of us have sat down and thought I need to do something here and you know your GPT or your Claudine just go Hi Seb, welcome, what do want to do? And you just go cool and tap away and then actually think No, I didn't mean that, need to say something else. Actually taking thirty seconds to prep some of the prompts you put in is a really good way of getting the prompts better, quicker, so that you don't have to go back and regenerate the whole time because it does make a huge difference. You also need to be really specific about the output. Don't just say, I want a presentation that, is this long. You have to say, what format you want it in, what tone you want it in, where it's gonna go, etcetera, etcetera. The more detailed and specific you can be about the output that you're looking for in a prompt, the better it is in terms of getting it right so quickly, getting it right first time prompting well. AI really needs to understand the why as well. Bad prompts don't really explain the why. Why am I doing this particular task? You have to give it context. It does mean sometimes feeding in a bit of stuff, bit of some documentation that always helps, but giving AI any context to why you want it to suddenly create a presentation or this particular document really, really does help for it to be more specific and takes far less time, to generate something which is really meaningful. The other thing that's really tempting to do obviously is to go no that prompt is wrong or to start again from scratch. But actually having the conversation people forget that the whole chat like nature of LLMs means that you can start iterating on the previous prompt rather than starting again with something completely new. That takes less compute and it's actually really easy to do. I think writing a better prompt is probably the easiest thing that all of us can do to help reduce the carbon footprint of the AI that we're using. The third one, and you might not be able to do this as well, but it's worth mentioning, it's worth just because you can ask some questions about this, is to have a look at your AI supply chain. This is really important. Someone told me about this the other day, and it really got me thinking. But most brands think their AI relationship is with the app on their screens. So whether it's Jasper or it's Claude or it's, ChantGPT, but it's a three layer supply chain, and kind of you're responsible for all of it if you really care about sustainability. So the app might be ChatGPT, it'll be on your screen, but it's behind that is their foundational model. So, for example, OpenAI, which is powering that and OpenAI is then using infrastructure like Amazon Web Servers to do all of the compute. So, this is a three level thing. So, as an organisation, what people need to start doing is looking along that three, layer supply chain. As a person sitting within an organisation or brand or anything, you can probably really only look at the first couple. It gets a bit messy there. You're not going to be, you're not going be able to necessarily change the infrastructure these things are based on, but it's worth thinking that that is the carbon footprint chain, of using AI. I want to share my current AI stack. This is always changing. I just having a chat with Joe before I came on, about the difficulty of setting up independent, AI platforms is that Google and Microsoft are putting so much money into their AI that actually, I think there is a concentration of all AI in those particular two particular organizations. But this is my one as a as a this is my stack as a creative person. Now I haven't gone for a small language model. I still use Claude, because it is really good for strategy and briefs and scripts and prompts. But if you read into their Anthropics sustainability initiatives, they're good. They're quite efficient about how much compute they put into each prompt and they give you guidance as well when they think that you're going kind of off piste that allows you to to spend less compute on it. So I do like Claude, and I do love what Anthropic stands for, but it's each their own on this. Adobe Firefly, I've started using over DALL E3, which was part of OpenAI. This is for image generation. They've got really clear science based climate targets. They've got renewable energy commitment, and they've got efficiency optimized interface as well, something that, DALL E3 didn't have. I do really like DALL E3. I miss it quite a lot. But again, this is just a more sustainable way of doing the things that I want to do with it. And you'll have your own things as well. Google Gemini, obviously, really good for for quick images and reasoning and research. Carbon Neutral since two thousand seven is pretty good. And, you know, they are Google are doing incredible things about, looking at renewable power data centers, as are some of the the other big AI players as well. You'll really be be reading a lot about that, in the news over the next eighteen months, whether they are wind and solar powered, whether they're nuclear powered. Unfortunately, there's a lot of gas powered renewable data centers being created in the US, but it's a bit of an anomaly. There is a lot of research going on, and that will, that will only accelerate. Runway, Gen four point five. I just still really like Runway for for generating hero films, and there's really good consistency. They get to the renders really quickly, and so the amount of compute they use is relatively low compared to some of their competitors. Canva, we all know Canva. That was mentioned at beginning of the presentation. Still love Canva. Think it's great. And you can do so many things on the platform that you don't have to have millions of others to do it. And then DaVinci Resolve. I used to use CapCut. I really love CapCut. It's a bite dance TikTok thing. So the supply chain is a little bit more murky on that. But DaVinci runs locally, so it is, materially better for the kind of the energy that it uses. As I said, my stack is always changing. I'm sure yours will be as well. Number four. So sustainable and AI is really, it's such a basic thing, but to ask more questions. Progress on sustainability issues, including AI, is really stifled by not initiating open discussions. They kind of get swept under the tar, you know, the carpet. It's really important to surface those and at least just ask questions about the provenance of, you know, the technology that we're using, whether we should keep optimising and looking at it. As I said, my my stack is, constantly changing. I guess yours should be too, because the AI is always changing very quickly as well. So, that's something that I would definitely, you know, ask you to think about and you know, we need education, and collaboration. This stuff is moving at such a fast pace. Anyone tells you they're experts in AI one minute will not be an expert, the next minute. I'm certainly not. And I'm trying to rapidly trying to to keep up as I'm working. But educate yourself, talk to other people, and collaborate with other people who you think share the same values. If you care about sustainability, you care about AI and you're excited but slightly taunted about it, then collaborate with those people. Find them. Seek them out. There are lots of them out there, not just the people throwing AI and saying it's the best thing ever, Not the people just saying AI AI is terrible. It's going kill the world. In the middle, where most of us sit, there are lots of people to collaborate with. So, try and do it if you can. The final thing really, the thing I just want to end on is to, you know, just use it responsibly and enjoy. AI is an amazing, amazing thing. It's awesome and it has but it has a really clear, you know, environmental impact. This is a journey that kind of all of us are on. As I said, I, you know, I don't think you have to be perfect on it. I'm I'm definitely not. But the real question is, you know, are you using AI intentionally at the right scale with the kind of right model? Are you willing to be kind of accountable for your footprint, either as an organization, or an individual? The best thing you can do really is just kind of track things and ask and keep tweaking. Insist on knowing how you are doing and your business is doing as AI is constantly, constantly evolving and kind of be relentless in that. Make sure that you check back on your on your AI. Make sure that you're keeping up with what's going on in the news, and then if you don't feel that, you are that what's happening is aligning with your values and where you want your yourself or your business to go, then you can switch things out pretty pretty quickly. I think the future can be sustainable if we want it to be, despite all of the headwinds that are there now. There's some amazing work going on, amazing innovation, which has been driven obviously by AI. There's some amazing innovation into energy use and storage and community benefit, which is also coming out of AI. It's not all negative. We all have the power working in advertising marketing to to do something better, I think, or to do what our job is better than we've done it before. I believe that we can all be part of it. Thank you very much for listening. I will stop doing this and come back to the I love it, Seth. Thank you so much. You can see in the chat you got folks like Roxy and Nicole saying claps, claps emojis. But also just there's some really interesting chat going on, throughout your presentation there, mate. So thank you so much for bringing this together. It's a good stuff which, really prompted people to sort of just be discussing it, I think that's exactly where we need to be at this conversation. I don't think we're looking for answers. I think we're looking for nuance and the right decision at the right time for the right context. Thank you, like sincerely. It's really, really appreciated. We've got some questions in the Q and A and so folks, if you've got any more, do pop them in to the Q and A feature, but I'll head straight to yours. And again, you know, the thing I like about today's session is that neither Seth nor I have sort of come in here pretending that Seth knows more about it than I do, but that we are scientists. What we are is marketers who are sort of looking to make the best out of it. So let's head into the Q and A set and see what we can do to make the best out of this conversation with the help of the community as well, by the way. I'd really, really value your comments in the chat. So let's go for the first one. It's quite a difficult question to answer, Seth, because the answer might be no. The answer might be that we have to seek this out. But do you know of anywhere that's reliable to find out information on energy consumption of different AI tools? And I guess where I broaden this out is like any AI tools or tasks associated. For example, a video generation is presumably going to develop need more water than say a prompt response. So how are you finding this out? Know what's interesting is that I don't think that you can go to a single resource on this and be really happy with it. So I've got kind of a Feedly dashboard which pulls in information from loads of different areas as well. Some of them from from, some Harvard or MIT, some of them from from Google searches which bring up, AI and sustainability, some of them from news feeds and you know Guardian, whoever it might be, and getting getting a broad view I think is the best thing. It's really really hard not to be not to find unbiased completely unbiased information out there. Every time I think I have done, I've realized that I'm leaning it's starting to lean left or right. This comes back to where the politics is at the moment. A lot of people are writing with a with a bias about AI, and I think the only thing to do is to set up multiple threads that allow you to look at one subject through multiple lenses. I know that's a bit of a cop out but that's what I would would be kind of saying at the moment. Nice. I I love that and I I find that a really helpful answer as well so like it's not a cop out at all because I think there is no answer. It's probably the best answer and it's, matched by Molly in the chat actually. He says, funnily enough, the big tech firms aren't transparent about energy use, so most of the data isn't available. So it's but what I can get from your answer there Seb, it's more about the principle, you know, and sort of like finding out, I don't want to go so far as to say the vibes, but like if you can get a sense of, you know, I'd prefer to use this for this reason, then that's possibly as good as we're going to get. Yeah. Yeah. Yeah. It's interesting, actually. And as I said, so I even just set a I set a Google thing to sustainability and AI, and it just pulls in every day. You can you can get to your point, you can get a vibe of what's happening at the moment. There is a big data center backlash happening in the states at the moment, because a lot of the data centers are being put in communities where resources are really scarce. And people look at it going, wait a second. What how is this benefiting the community? Having one hundred people build this is great, but then it can run itself. And actually, it's just we're having energy spikes and price spikes all around it because it's taking because it's taking too much. So, yeah, I think get regular alerts going and just keep yourself up to date on it. Would be the best thing. Nice. Love it. Thank you very much mate. Appreciate that. Let's take the next one from a very brilliant, Sarah. Sarah is just a wonderful human being. So I'm not surprised about the spirit of this question coming from Sarah, which is, what's a good way to manage or persuade colleagues who tend to use AI for everything to do what they do to be more intentional about it. So I think, by the very nature of folks showing up today, these are folks who are interested or curious, but there will be people who, for one of the better phrase, will bow at the altar of productivity and more and more. How do you get to that? That is a big naughty question. It kind of depends on the organization, the colleagues you're talking about, the level they're at, etcetera, etcetera. But I think I I think maybe I could reframe it in what what is not good to do. And I think, criticizing people for their the way that they use AI is not good. Ai is kind of AI is bad and you shouldn't be doing it type of thing is really not a good way as well. So, I think, certainly when when I've had conversations like that, being enthusiastic about AI, but wanting it to be optimised for the in the best way it possibly can be so there's not you know words like wastage and optimisation, often appeal to people as well and doing it really efficiently doing it in a way which is like, streamlined all of these good things I think is having positive language which doesn't mean just it doesn't feel wasteful it doesn't feel like it feels much more focused and I think certainly those are the conversations I've had God I love IR isn't it brilliant it could be even better if Yeah nice I love that and meeting people where they are it's a it's I guess it's almost marketing what I want to assess it so isn't it? Yeah that's fab thank you very much for that and I love that spin on it And by the way, I appreciate how you've shown up in general, which is not so much, you know, like good or bad, but you know, just sort of seeing it as it is. Like, it's exactly what I hoped for from this. So I'm very grateful for that as well. There's two questions that have actually come in on the very same topic. So one is from Sarah, again, who's on fire with great questions. So I'm going to take this one instead from Nicole, which is in the same spirit. It's a little bit like the first question that we spoke about when it came to, sort of knowing what is doing more of consumption elements of things. But Nicole is asking, have you found any easy ways to track how you're using or consuming data or AI and reporting that back to your business? So I think in your first answer, we acknowledge that some of the data isn't available. So actually being able to give like a decisive answer is, I would say, not an impossible. But in terms of that reporting element, how are you gonna? Do you do you know what? I think there is no standardization yet, and and so we are really in the early stage of things. What I did, I remember I I recut the Guinness commercial, years ago. Well, not years ago, last year, with some new footage about kind of environmental, you know, sea pollution. And what I did is I wrote a, a workflow of how many prompts I'd put in to, in order to get there. And then the amount of prompts I used to a film I created last week was half of that. And so I'd realised that I'd made some progress and I think what you can track is progress. What you can't track really well at the moment is actual specific energy stats. But if you are tracking progress and you realise you're getting better at prompting, you're getting better and more efficient at creating the pieces of content you're going to do, that is a good thing. I think you have to actually do it manually at the moment, which is how many prompts have gone into this particular piece of work I've done, whether it's a piece of video, whether it's a presentation, whether it's anything else. And keeping track of that, manage like a diary. It's like a usage diary. If it's moving in the right direction, so you're doing less of them, that is a good thing. I know that that again feels very analogue kind of these days and there will come a point where you can actually measure amounts of energy. But what we're trying to measure is progress. Because progress is a great motivator for doing more as well. So if you think actually God last year I was like this. This year, I've got much more efficient to this. That is a good thing. That is, you know, spending less energy and it means they're moving in the right direction. All of the tools to measure actual energy will come but they're just not there yet. It's it's it's fascinating. And so I'm combining a thread in my mind of the earlier conversation with something I read today which was around a phenomenon in Silicon Valley called token maxing And so the theory of token maxing for those not familiar and I wasn't until today was the more tokens you consume, the better it will appear to your company. So one step further up, you know, when you use specific AIs, you use tokens to manage that interaction. And some companies such as Meta, and it might just be one employee. I don't need to sort of portray this as like a company wide thing. But one employee built a leadership board, which sort of said this person is using the most tokens and therefore is the most productive person in the Now, what I find really interesting about what you've just been speaking about there Seb is like actually using less tokens is the progress. Yeah. And that's like this cultural thing, you know, where we bow at the alter of productivity, and therefore more is more. But somehow, a part of the narrative is about switching that and sort of saying, less is more less is more efficient than more is more efficient if that makes sense less and better yeah it would be a really interesting thing wouldn't it to do a test within a company where you all have to do the same task as efficiently as possible using AI. I think that would be a great thing to for people to return to time again. It's like, so and so did it in using fifteen prompts and this that and that and you know, someone else got the same result in six. And then so what can we learn from that? What was the difference in that? And I think God, there's there's whole internal programs you could do with that could be amazing. And, apologies also to Deborah, in the chat who rightly points out that I should have been saying fewer the entire time there as well. So But no, I do agree with you Seb. Thank you very much for speaking to that. Let's take a quick look at Darren because Darren's asked an interesting question because of course the framing of today is around AI, and AI is new. But Darren says, I personally didn't see much conversation around sustainability pre AI, around normal Google, Bing, Firefox, etcetera searches. Of course, they're not apples and apples, but I think the point sort of still sustains. How does the impact of AI search compare to legacy search, including website page clicks, and engagement? And again, you know, we're speaking definitives here in a world where we don't have definitives, but I wonder whether you have a sense, Seb, of how we can compare what we used to do to what we're doing now. Well, we can be sure that we're spending a lot more energy now because the AI is still being powered by large language models. The return on AI searches is just like a prompt. It just happens to be in a different format. That's what they're doing. So, rather than re engineer stuff at the back, as far as I'm aware, What they've done is just surfaced it in a different way. So, instead of you going in and having the chat, in, you know, within GPT or Claude, you're doing it within Google and it is still drawing on same large language models to give you the answers. So, the answer is it's just the same as using AI, but just in a different format as far as I know. And I've again, I've read nothing that suggests those have been super optimized and clearly versus old search. That's that's night and day. I mean, there there were there were some stats actually knocking around which are a bit out of date now. I kind of had one that's bad. It was about, you know, a single chat gbt prompt uses, ten times, what an LED light bulb does. And there was a, sorry, forty percent more than a Google search that we got here. So, ten times more, but forty percent more now. And I think this is this is the point. All of the energy usage which is horrendously large at the moment is coming down. It is coming down because so much time and effort and money has been put into, how do we reduce this because it's the one thing which will hold AI up in the future But to come back to Darren's question, normal search versus, AI search is night and day, and AI search is at least forty percent more energy, sucking than the old search was. Thank you, sir. I'm gonna, I'm gonna, prompt this for the better word with you, because you mentioned your Guinness ad earlier and I don't know whether you've had that primed before we went I haven't, annoyingly. Haven't got it on me, sorry. No, that's all fine. So what I'm going to do is I'll find it while you're answering the next question. So I'll describe it verbally for folks that you, as you said, you you recut the famous, horses in the waves advert, with environmental message and and you did that using AI. Halfway through your talk, you sort of presented a slide which said to sort of start with the end in mind, which for me was the big takeaway here. And I wondered in that specific instance, here you have the opportunity to use a video, and create a video which has an environmental message and therefore is a positive thing, but you're using AI to do it. How do you what's your sort of thought process as you're going through that to sort of like, oh, this is justifiably worth the effort in terms of that creation, etc. Yeah. It's a reward, you know, yin and yang thing, isn't it? I think, look, I think my my point about, I guess, the Guinness Surfer, one of the most iconic pieces of advertising ever made, was that, we spend a lot of energy and time creating new content all the time. I mean, AI has allowed us to be very, very productive. The world doesn't need more content in it. It needs better content or better focused content. The worry I think about AI is the bland blandization, if you could call it that, of creative, where there's just so much stuff. The pipe has just become even wider and it's just pouring stuff out all the time. My my thinking with the with the Guinness surfer was the people who remember that ad from the first time around, it was very iconic, golden age of kind of, you know, big high budget, really expensive to produce advertising, are now going to be at the top of lots of organizations. There'll be CFOs, there'll be CEOs, they'll recognize that stuff. Is it possible to take iconic content which hasn't run its course in terms of its impact and creativity and reimagine it, recycle it, if you like, so that you're not having to produce a whole load of new stuff? And look, using a few prompts and inserting some imagery is a tiny amount of energy compared to the potential good that you could do to spread a message, about not polluting our seas, for example. So, for me on that one, it was a it was a kind of risk reward or an open fact, in a new film I've been doing about the roller coaster of energy usage, as I was sort of chatting to you about before, that would have involved a three day shoot, enormous amount of trucks and lights and everything else which would have been a high, a high carbon thing. We all know how much, you know, carbon, you know, physical shoots do. At the same time, it would have employed thirty people. So, it's a constant, it's a constant balance and one the reason that the heart was torn a little bit in there is because I just don't know where it all lands necessarily. I really want to use this stuff because I'm full of ideas and I really really want to get those ideas out into the world and hopefully they can make a positive impact at the same time being aware of there are impacts and downsides. On that particular one I thought there was more upside than down. I think that final sentence is probably where we're at to a certain extent you know there's that final sentence of you know like there's the risk and reward but you know in this instance I thought the reward was better than the risk is you know, if we can leave someone with that in mind today, I think we're kind of doing our job to a certain extent and I'm not really sure that I held that mindset coming into today's session. You've changed at least one person's mindset and I hope there'll be a few more in the chat as well. So you know we've got Natalie as well saying here in the chat thank you for raising awareness about that. It's a bit overwhelming to think about where this could lead for the environment. It can't be be policed easily at all. And I think overwhelm is a big thing but I think there was another chat comment earlier about like I'm avoiding using AI for all these reasons and I think that in itself is not beneficial either so I think I think you may did you know words that someone I remember stressing about this massively when I was doing that Cambridge course and someone said to me just do what you can It's like okay that was really to the pressure off. It's like you don't have to be amazingly perfect about things and you shouldn't really be pulling anyone else up on stuff too but do what you can. If you care about it then take an interest in it and and do what what you are able to either within your means or your seniority or your ability to do stuff. And you know, sometimes this stuff costs money as well. You know there's subscriptions and this and that and the others. Know just do what you can. If you can do what you can and you can feel good about the progress that you're making comes back to this is where I was, this is where I've got to, then I think that's a good thing. Thank you mate. I think that probably speaks to something else which we haven't covered in today's session but does feel relevant which is that do what you can also speaks to those in a context where they're working in an organisation where it feels like AI is inflicting upon them or there are, there are things which they can't control which maybe they're being asked to do which they're uncomfortable to do because of those environmental reasons and and yeah and so for you to sort of acknowledge that feels quite important. Yeah and I think you know there are there are options you know there are options as we said you know you the the smaller language models the ability to look at how you're using them. Those are things that that are doable for for hopefully doable for anyone, you know. So, yeah, small practical steps, not doesn't have to be a huge, great leap. Lovely. Thank you, sir. Mate, I I feel like with the questions that we've got remaining but also the stuff in our head, my head that's left to be to ask. It's it's stuff to a certain extent which has already been covered in today's session and I I don't want to sort of flock the horse to speak and and just chat for another eight minutes just because, just because we can. So with your permission, like, I I guess I'll with one thing and and that's just asking if there's anything left in your head on this discussion which we haven't covered today where you're like you know what whether you know it's worth sort of mentioning this thing. Not off the top of my head. I think I think just that if you're wondering about it and you're feeling kind of frustrated or not knowing what to do, I don't think anything has to be binary at this point. You don't have to not use AI. And in fact, I would not suggest that you go down that route. AI is not going anywhere soon. So but also, don't worry about the rabid. You must know everything now, sort of camp, which I don't think is very helpful either because it makes us all feel inadequate and that we should be up all night learning about this stuff. A lot of it will come kind of by osmosis. I think where you can just take a bit more of a lead is in being curious and asking questions, and adjusting your own moving to your own path and your own beat and I think you know being happy with the progress that you're making. I love that mate and I think that that spirit of you are not behind is something which applies so strongly across all of this topic in general, know, never mind just the the sustainability angle. I just feel like, an overwhelming sense of gratitude really for you, Seth. Like, I I know you're outside of TMM and so, like, I think an awful lot of you as a person and so for you to show up as you have for the community today is fabulous and folks if you're watching today and you'd like to say thank you to Seb in my mind I know that he wouldn't necessarily ask for that, but in my mind, I just really appreciate the effort. So thank you Seb and thank you for Thank you, I really enjoyed it. And, yeah, and if you want to chat about any of this stuff, then you can find me on LinkedIn somewhere. It's a symbolist. I've just popped it into the chat feature so hopefully you'll get many people saying thank you. Thank you all for being completely wonderful. Thank you Seb. Thank you also to to to Canva by the way. I I think with all of us we won't be able to do this stuff we're going have these chats if it wasn't for those companies so they're supportive of this sort of stuff so thank you to them as well. With all that said we'll hopefully we'll be back in two weeks We'll be back in two weeks with a very different session and so we hope to see you there. But thank you all very, very much. Thank you Seb and we'll see you. Cheers everyone. Bye.