EP 312: Is AI only a resource hog? A greener side to AI

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business and everyday life.

When looking at artificial intelligence, I feel sometimes we only look at what it can accomplish, but not what it costs or what we should be doing as business leaders to maybe offset that cost. Because here's the reality. When we're using generative AI, when we're using large language models, it's extremely resource heavy. So much more so than, you know, doing other things that you may be doing on a day to day basis on your computer. So I'm excited today to have a special guest on where we're going to be talking about. Is AI just a resource hog or is there a greener side to it? So we're going to be talking through some exciting use cases on how this organization is using AI to what I think to create a very greener future. All right, I'm excited for today's episode. If you're listening on the podcast, as always, make sure to check your show notes. You can click and go to your everydayai.com Sign up for our free daily newsletter. We will be recapping today's show in great detail as we do every single day, written by me, a human. And normally we do this live and we go over the AI news, but our guest today is from Australia. So, you know, normally at that time it would be, I don't know, like 1am or something like that. But we will still have the AI news and everything else that you love in today's newsletter. So make sure you go check that out. All right? With that, I'm excited to bring on our guest today to talk about this other side of AI, the greener side, and how they're using AI to really do more impactful work and hopefully inspire some of you all to think about using AI in the same way. All right, so please help me welcome to the show. There we have him, Emerick Modu, who is the founder and CEO of Lord of the Trees. Emric, thank you so much for joining the Everyday AI Show.

Thanks, Jordan. Hi, everyone.

All right. And hey, Emorett, can you tell us a little bit about what you all do at Lord of the Trees?

Yep. So in a nutshell, we use drones, robotics and AI to restore biodiversity. So, you know, mainly plant trees all around the world. So the aim, really the mission is to heal the planet or, you know, plant more trees at a faster rate than we destroy it.

And I think that's a Great mission. And we'll talk about that more later because I think when it comes to AI, there's definitely a side that people don't look at. But before we get into that, Emerick, I'd really love to hear just maybe some examples of even how you are using AI. Right. Like when we talk about planting more trees, I think a cause we should all be paying more. More attention to. But in my mind, I think of it as a very laborious process. Right. Like going out there, you know, probably having groups of people walking around, finding the best places. How do you all use AI to make that process maybe a little more efficient?

Yep. So the way we work is really divided into four phases. We on phase one, we do landscape analysis. So depending on the size of the terrain that we have to work on, it could be anything. It could be rehabilitating a mine that's closing down. It could be assisting governments after bushfires, wildfires, and we're starting now to work with farmers as well. So depending on the size of the terrain and of course, the accessibility, we either send a drone, so the pilot, the co pilot, and then a little array of drones, or we use satellites. The data that we collect is then processed. So that's stage two. And that's when AI comes into play where we look at the data that we've collected into the landscape and we match that with the plant requirements. So each species, as you know, has different needs and ones. Some plants like sun, some likes a bit more shade, some likes to be in drier parts, as opposed to others that like very moist environments. Some are not susceptible to wind, etc. Etc. We look at soil quality as well. Some like acidic versus alkaline soil. And based on these data that we have on the landscape, we match that with the plant requirements and we create what we call a digital twin of that landscape. And AI helps us to come up with what is, in theory, the best planting plan for this particular. This particular landscape. Once we have that into place, we use AI again to help us calculate with precision how many seeds of each species will be required for this particular job. We then go on onto phase three, three, which is the planting. There is two things that happen here for temperate forest. We create what we call seed pods, which is it's all done in a lab. Each little seed is isolated and we make a little backpack of nutrients to give it the best chances of survival and better chances of germination while it's dropped into the landscape. We load those seed bulbs onto a drone, then we fly the drone into the landscape and based the result from the AI. The drone, which has been pre programmed, can shoot with precision. So we do precision planting. We fly the drone no more than 3 meters above ground, which is 7 to 8ft maximum. So we, we're not going too high in the sky when we do that, just because we want to make sure that the seed pods are in the zones that have been identified for the best chances of survival. And that use of drone technology and robots enables us to cut the time spent by about 13. So we can work, in essence, we work, we're able to work 13 times faster than if this was done by hand on the back of a tractor using traditional forestry methods. Once this is done and the drone, as you can imagine, can access difficult to reach terrains. We work on all weather conditions as well. The drones don't take breaks and they work at night as well, which is great. When we have big landscapes to reforest and after the planting is done, we move on to the last phase, which is the landscape monitoring. This is the phase that takes the longest when we build carbon credit projects, we are there, we are going to do the monitoring for the next 20 to 30 years. So usually 20 years in landscape environments, 30 years in more temperate forest. And when we work on biodiversity projects, that's a 99 years time frame that we're looking at. So yeah, that's really us, that's how we use AI in our operations.

So now, now I have so many questions. Right, so, so let's start here because that was a lot to unpack all at once. So you know, Emerick, you, you, you talked about kind of the different phases or different stages that you're using artificial intelligence to just grow more trees. Right. But you know, can you maybe walk us through how it actually works? Right. So even like AI, right, when we, when we use AI, sometimes it just becomes a buzzword. So you know, are you using, you know, traditional artificial intelligence, you know, deep learning, machine learning algorithms you're building in house, or are you using like as an example, you know, APIs of common large language models that we use daily, like on a daily basis. Right. You know, from OpenAI, Gemini, Claude, et cetera. So yeah, walk us through a little bit more of the technical side and also, you know, what that has led to in terms of just more efficiency and productivity.

Yeah, so to answer your question, it's really all of the above. It's quite extraordinary. So when we do landscape analysis or the monitoring of the landscape, we collect as we fly the drone. I'm going to stay with the drones because we use drones and satellites. But let's keep it simple. Let's, let's stay with the drones. We use for landscape analysis a LIDAR component that's attached to the drone and then we fly the drone. We collect 2,000 data points per second. So you can imagine the enormous amount of data after an hour of flying that we collect. This data is stored into servers that are scattered all around the world. And after that we develop our own algorithms. So we use Python for that. It's all in house built and we add layers and think of them as. There's a layer for each plant species that we use, as I explained earlier, but there is also external layer, we call them external layers. When it comes to biodiversity, we have a very unique way to engage and look at landscapes that we work on, which is a very holistic way to, to work where we actually, instead of zooming in, which we've already done with the landscape analysis, we actually zoom out and we need to understand what is around and what is going to impact the landscape. Working with nature is always like, it's not something that is static, it's always moving. There is always something happening. There is not one day which looks like the day before. Every day is unique. The weather can change from one day to the next. Species move into the landscape at any time and we take that into account. So just to give you an example, related again to AI, when we look at biodiversity that can help us with mainly focus on birds and small mammals that are going to help us as we regenerate the landscape, that are going to help us to naturally disperse their own seeds into the landscape. And when it comes to soil analysis and soil biodiversity, we are very, we look a lot at fungi and how fungi is already present in the particular landscape that we work on, but also around the landscape. And that's very important, especially when we work on grounds that have been severely burned. Three years ago, I'm sure you've heard about the big fires that surrounded Sydney for more than five months. And those fires, same as in California recently, were so intense that they've actually completely sterilized the soil. And this, they completely burn the natural seed banks occurring in the soil, but they also completely decimated the fungi that helps in creating the mother forest, which is a healthy soil for those environments. So we look at all of that and AI helps us to, yeah with creating scenarios that we layer on top of that digital twin help us to make decisions or actually change A few things that we thought might work, but actually turns out that it's probably, there's probably a better way to look at that. So it's a very interesting way. Now when it comes to, so that's on the reforestation side. When it comes to the tech side of things, as I said, we use. So there's two elements that work.

In.

Perfect symbiosis in our operations. You have the high tech. So it's obviously the technology, the drones, the robots and AI that work in harmony with low tech and low tech is, so it's high tech and low tech. And low tech. It's a term that was created by an amazing person. Her name is Julia Watson. It's spelled L O T E K and it stands for local traditional ecological knowledge. And it is all that ancestral knowledge of the land and the forest which is kept by the forest people, the native indigenous people that we work with, which is something that AI doesn't know. So machine learning comes into play in that regard. So I give you a few examples when it comes to understanding the landscape in that regards, we look at plants that are, that have medicinal or cultural value that those forest people would use for cultural practices. We look as well at the fire management, which is a big, it's very unique in the case of Australia to have regular fire regimes in order to help the landscape. This has been done for, you know, thousands, thousands of years. I know there are different policies around the world. I used to live for four years in California and I remember Smokey Bear, you know, as soon as he said, now the problem that you have with that is when is you have an accumulation of litter onto the ground. So that the day you have fires like the ones that occurred last year and the year before in California, you have an incredible amount of litter that is just fuel and that is creating even bigger fires. So that's one thing that we take into account. And the last one is anything, the ancestral knowledge that surrounds the collecting of the seeds.

Yeah. And I think Emerick, like that combination that you just talked about there, that kind of high tech, low tech. I think it's extremely important, Right. Even if we're thinking in, in a corporate setting, right, to still have the high tech of generative AI, but you still have to have that low tech, right. Or that local knowledge of your people. Right. And I, I, I think it's important to strike that balance. And, and going back to something you said there, I did some rough, rough Math, right. So 2,000 data points per second an hour of Flying, that's like more than 7 million data points. So you have a lot of data, you're using a lot of AI and that's resource heavy. Right. But at the same time, you guys are literally making the world greener. So before we transition, I do want to ask you like just overall, how many trees are you planting or how many acres are you able to reforest a year using kind of this high tech, low tech sweet spot combination? Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on gen AI. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead of and some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you can partner with us too. Just like some of the biggest companies in the world do. Go to your everydayai.com partner to get in contact with our team or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and built a straight path to ROI on Genai.

Yeah, so when it comes to, I mean, that's a good point. I want to go back to your first point, which is the energy resources that we use, we are very aware of that. And when you look at what we do, you can't really, even when you have a service, it could be any service or any product, the end product or the end service that you provide cannot be looked at on its own. You really need to have what's called a full picture of the life cycle of your entire operations from anything that is being created with the tools that you are using. In our case, AI plays a big part of it. You mentioned the enormous amount of data that we collect that is stored in those data centers around the world. So we are very aware of that and it's part of our DNA, obviously to use AI for good and to use drones for goods and to do good for the world. So we're very aware of that and we monitor all of those emissions and the weight of the resources of the additional resources that we put on the planet to actually sustain our operations. Now, we're very lucky in the sense that, yes, we do plant our own trees. So to go back to your second quick, to come back to your second questions. The swarm of four drones in 48 hours can plant a million tree, a million seed pods. So that's 48 hours. We allow four hours total out of those 48 hours for battery and seed refill. But on and all those four drones can work tirelessly for 48 hours and plan 2 million seed points. Now, we recently became a B corp, maybe a month ago and we. So we record our. We have a very good understanding of our emissions, not just when we travel to go from point A to point B with the team, but on the resources as well that we put, especially when it comes to electricity because that's many, you know, AI being powered by, you know, with energy and we plant more trees than we consume. I'm going to say that. Yeah. Than the energy that we consume. So we're lucky in that sense. And I'm very. It's something that is very ingrained in the DNA of our company. It's part of our culture as well, the company culture. And for everyone listening to the show today, I think we have a duty as, you know, custodians of this planet to do the best we can. It's not, you know, it doesn't have to be complicated, you know, planting trees or being supporting projects like the ones that we have that actually do good for the planet and create more positives than negatives are a big part of that. That mentality that we all need to get as a collective in order to make. Yeah. In order to keep flourishing as a species.

Yeah. You know, Emerick, again, great, great points you bring up there because yes, you're using a lot of AI, but you're obviously literally making the world greener and you know, the ability to plant millions of trees in a week, which is, which is pretty fascinating, but, you know, not everyone's in that position. Right, but. And the use of AI is, is just going up. A Morgan Stanley report recently said generative AI's power demands is expected to grow 70% annually. You know, another study that came out showed that a simple chat GPT query, you know, requires 10 times more power than even a simple Google search. Right. So I know there's no easy solution to that. Right. It's like, okay, everyone's going to be using AI more yet there is this kind of almost ugly underbelly of it. So for companies out there, for business Leaders that aren't in your position. Right. Who aren't able to plant millions of trees. What are some of the decision making processes or what important questions do business leaders need to keep in mind when it comes to their generative AI use?

Yeah, so I think you need to, yeah, look at, I mean that, that's exactly what you say. It's to look at where you stand and where you want to actually move forward when it comes to, you know, integrating better management practices. Right from, you know, the top down. I think that's very important to have the support of, support of top management. It's really a top down approach. We do work a lot with governments and tier one companies in that regards that get behind us to support reforestation projects all around the world. So it's not just reforestation project of just planting trees for the sake of generating carbon credits that happens, but we have projects as well that sustain endangered species. We do work at the moment we have a mission happening in Borneo as we speak to, to work with a great organization called uff, the Orangutan Republic foundation of Borneo. We do work in Vanuatu and on a project that is mapping, helping the government to map and protect their mangroves. And we have, as a result of that we're able to generate biodiversity credits that specifically focus on turtles. Just to give you an example, another project in Mexico coming up to do with three endangered species of bats that migrate from the south of Mexico all the way to south of America. So it's varied and there are avenues or ways to basically compensate and you don't necessarily need to have millions in the bank account to make that happen. It all starts with, you know, small steps to. I think the very first step is the awareness of what you just brought. And the second step is a commitment to create positive change. Whether, you know, you back those projects or you decide to use green energy or be behind other projects. We're not the silver bullet of environmental issues. There is hundreds of companies that are doing great work around us that we work with from, you know, small farmers in South America to bigger corporations in other parts of the world. But I think it's the collective, you know, the, the collective of those little actions and everyone is doing their bit to. Yeah, yeah.

And I think maybe another example going back to two points that you talked about earlier is. Well, when you use generative AI, even if it's heavy use, even if you're gobbling through millions of data points, ultimately you are probably saving time in other places or you're probably saving compute or just human hours. So going back to your original point, using the drones and AI and robots, it's 13 times faster. Right. Than you would with humans. And there's obviously environmental costs of humans walking around and, you know, the energy that they need. Right. I'm sure you could go on and on, but, you know, one other thing that you talked about that I want to dive into a little bit deeper is this concept of digital twins. So we've talked about it here on the Everyday AI show before. You know, I was lucky enough to be out at the Nvidia GTC conference this year talking with a lot of the individuals who are building the future of digital twins. But maybe Emre, just give us a quick overview of what the heck is a digital twin and maybe give us a concrete, quick example of how you're using digital twin scenarios.

Yeah, the first one I've already covered when I was talking about our operations, when we do landscape analysis, basically, to put it in very simple placement terms, digital twin is the representation of the landscape onto your computer in a 3D thing that you can rotate and look at and turn around a little bit. Like think of Google Map. Right. That would be the best of Google Earth. That would be the best example of a digital twin that we have of the particular landscape. When it comes to tech, we use digital twins as well for our tech. And I'm going to be, I'm going to give you another example here where we are building new drones. In part, we have a joint venture with the University of Queensland to build a new drone for them by the end of this year. And we've built the drone that came up. So it's a prototype at this stage. It flies really well and it does all kinds of really cool things. But what we've decided to do, because we need to fly this drone in different weather conditions, instead of us waiting for those weather conditions to happen, and then take the drone out, take it in, take it out again, and verify and do all kinds of testing, which would take us literally a year.

Right.

Because we want to test that drone in extreme conditions as well. So when we are hit, for example, with a cyclone, when the weather gets a bit insane out there, and compare that with when we have actually better weather conditions, we have put, we have put the CAD plan into specific software that our engineers are working with, and the software has been able to look at different scenarios of pressure and wind condition that mainly impact in our industry, you know, the flying of the drones. And based on the results we are, which happen in minutes, we've been able to modify the building plan of that particular job. It would have taken us, as you said, a year to actually come to those conclusions and do a lot of, you know, trials, but within a few hours and yeah, it was just incredible to see what would happen when it comes to velocity, pressure, thrust, etc. Of, of that drone just on a computer. So that's when using AI for good really speeds up the operation, reduces cost and enables us then to be back outside on the ground or in the air as soon as possible so that we can, you know, we can plant more trees and we can do more things.

Yeah. And Emerick, we've talked about a lot in today's conversation. This is going to be a jam packed newsletter, by the way. But, you know, as we, you know, wind down for this episode, I mean, we've talked about a lot, you know, AI is increasing energy consumption and the environmental impact, offsetting, you know, how companies maybe should be offsetting their AI carbon footprint using digital twins. Right. Like you just talked about using digital twins for landscape analysis and energy saving. But maybe what's the one most important takeaway that you want other business leaders to glean from this conversation on maybe creating a greener side of AI?

Look, I don't necessarily think it's the greener side of AI. I think that should come naturally. I think we'll get there. I mean, you're right, the AI is growing. I don't think growing is the right time. I think it's exponentially growing. Like when I look at, you know, what the algorithms now spit out compared to just even like a month ago, I'm absolutely baffled. In a good way. But as I said earlier, if there's just one takeaway, I would like your audience to stay with. It's just one step at a time. It says take the first step. It doesn't have to be daunting. There are people out there that are doing the right thing, still doing great things. We're more than happy to help and talk to people if they have any questions or how could that happen? But yeah, just one step at a time.

All right. Hey, well, Emerick, thank you so much for your time and joining the Everyday AI Show. We really appreciate it.

Pleasure. Thank you. And I love your show, Jonathan.

It's great to. Speaking of one step at a time, the next step that you need to take is going to our website, your everydayai.com. sign up for that free daily newsletter. We will be recapping everything that Emmerich talked about and a lot more. So if this sparked your curiosity and maybe you want to do better understand of, you know, better understand different ways that you could be using AI or, you know, some of these projects that Emmerich talked about, it will be in the newsletter. So thank you for tuning in. Thank you for your time. Thank you for joining us. Please. We'll see you tomorrow and every day for more Everyday AI. Thanks, y'.

All.

Thanks, everyone.

And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit your everydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.