AI and Water Usage: Do we really need to be worried about it?

0:00 AI is revolutionizing everything from climate modeling to social media, but it's still creating a problem, a new kind of environmental footprint, let's say, that we rarely talk about, and we're gonna talk about it today about water use. What if data centers powering AI are quietly reshaping the way water moves through our world and maybe even affecting the ocean life itself? We're gonna talk about that on today's episode of the How to Protect the Ocean Podcast. Let's start the show. Hey everybody. Welcome back to another exciting episode of the How to Protect the Ocean Podcast. I'm your host, Andrew Lewin. This is the podcast where you find out what's happening with the ocean, how you can speak up for the ocean, and what you can do to live for a better ocean.

0:43 By taking action. And on today's episode, we are going to be talking about AI. Now, I have to admit, I am a big user of artificial intelligence. I have an account with ChatGPT the $20 account a month, and I use it to work this podcast. I do this podcast by myself. I publish three episodes a week. I work a full-time job. I have kids, I have a family, you know, that I will like to spend time with and I like to enjoy my evenings. So instead of spending an hour and a half or two hours to like do the show notes myself and write it out and go through all the process and you know, mark everything, I use AI. I use AI to come up with the basic things based on my transcript, which is using AI to get a transcript out.

1:26 And then I publish it. I put it up, you know, on YouTube, on the podcast apps everywhere. and it helps me. It saves me at least two hours. When I do interviews. I use AI to edit the interviews, and that helps me just kind of like go through a process. It takes me two minutes to edit an interview instead of, you know, it takes me two hours. So it saves me a ton of time. I do really appreciate it. It allows me to create more content through video and audio like I do on my podcast, which I really, really enjoy. The other stuff in writing is not my forte. It's not something that I thoroughly enjoy doing all the time, although I like to get my message across. I'm starting to enjoy writing a little bit more as I start to journal each and every day.

2:07 There's times where I use AI to really help me get through the process, and it helps. However, with that said, I have recently realized the impact that AI has on water usage and why are we talking water usage? Well, this is how to protect the ocean. We're talking about fresh water. We're gonna be talking about oceans as well. And the effect of AI to give you a broad understanding of the effect of AI on water usage and really to set the stage is, let's be honest, since Open AI came out with chat GPT, and even before that, it's kind of scaled to a point where I don't think any of us really predicted. Maybe a few of us really predicted, but I did not predict how big it was gonna be part of our lives and how quickly it was gonna be part of a live, almost to the point where it's getting creepy.

2:54 We're getting people who are building robots that become girlfriends or partners to humans. So like robots that are partners with humans, we're seeing that happen in front of our eyes. It's kind of weird. It's kind of creepy, but we're seeing it. It's gotten to the level where people are using it to write their emails. People are using it, you know, as a search engine. People use it. Like for me, I get my research from it, not just like, Hey, ChatGPT, give me your opinion on this. say, pull all of the resources you can together to get me the research that I need, including journal articles and stuff like that. And it's great at doing that. Now, when I was researching this episode and doing AI and water usage, I was a little concerned that chat GPT wouldn't give me the resources that would really like talk about the information that I need in terms of how much water is being used per year and all this kind of stuff.

3:45 I actually use Google and other search engines. And other places and looked at reports to get some of my information. and it was pretty much the same. So it was pretty good that way. I was pretty happy, but it's definitely weird using AI so much in our lives. It's happened so quickly especially in my life, I use it quite a bit. I talk to it when I go through my strategies of what I want to do next. By the way, on YouTube, I've got my first level of monetization. So I can do memberships, I can do merchandise, which is coming. I'm gonna include that on my Patreon as well for people who don't watch this on YouTube and listen, but the idea is like really this. AI is a productivity tool. It helps you with productivity to become more productive in less time, which really helps a lot of us.

4:26 But the thing is, is with this boom and this huge scalability of ai, we are seeing data centers being built all around the world, and these data centers are able to handle all the inquiries that chat, GPT and Microsoft. So there's all these different types. Copilot is for Microsoft that we're allowed to use at work. And there's all these other different ones. There's Gemini with Google. They all have it. And every time you make an inquiry, it uses water and you're like, Andrew, how does that work? I mean, robots aren't thirsty. Computers aren't thirsty. They're thirsty in a way though. They're thirsty in a way. and I'm gonna explain it to you. And there's two ways that data centers that are being built to handle the amount of inquiries that are coming through.

5:10 Like we're looking at billions and billions and billions that are coming through. What happens is computers will take those inquiries and they will generate an answer and they've gotta do all the research and all the backend stuff. And they've had these servers that are all situated within these data centers. Right? And these data centers look like massive warehouses in certain situations, and sometimes smaller, depending on where they are, let's just picture them like a big warehouse, like almost like an Amazon warehouse. And you have all these like servers in there and they're doing all these every time an inquiry is made. what's the weather gonna be like? Or, you know, help me write these show notes. Or, how many species of bees are there? Like, all these different questions, like, help me do this, help me do that. It generates heat every time the computer is processing, right? Because like you've seen laptops and you've seen computers, and you've seen TVs every time you use it for a long time.

5:55 If you feel the fan or the back, there's fans in there to cool that thing down. And so a lot of the times when these data centers and the computers are working, they need to be cooled down because they're heating up like crazy. If they heat up too much, they'll just shut down. So they need to be cooled. They use water to cool that, hence the water usage. So a lot of the times right now, from what I know, and this is very rare, where you see where they have recycled water, but they're actually using a lot of potable water, like drinking water. like drinking water sources to take the water from lakes and rivers and wherever they can and bring it through their computer system to cool down the data center, right, to cool down the servers that are working constantly and constantly and constantly.

6:35 So now you have that part of it where they're being cooled and all these pipes are running through with cooled water and it cools everything around it. Then you have, how do these data centers operate? They need electricity. To get electricity, they need to like heat water, and they need to like get water in. So a lot of times they're being consumed by the heating of water to get that energy. So like hydropower and all that kind of stuff, that's what comes through. So they're using that type of water to power the electricity. They're also using fossil fuels. Everything like that, which we'll talk about in a minute. But essentially that's what's happening. So you have direct cooling and you have indirect cooling through electricity generation for that type of water use. So that's essentially what's happening. Now, here's the big problem, is they have to find places to build these things across the world, not just in North America, but across the world, right?

7:21 So where they go, where do you think they're gonna put these big data centers? Are you gonna think they're gonna put a downtown Toronto or downtown New York City? No, they may put one or two, but they're gonna go where the land is cheap. They're gonna go where the land is cheap and where the land is cheap usually is not the best area. There's not a lot of water already. Could be in poorer neighborhoods or poorer areas like where the people are poor, they don't have a lot of resources. They're putting it in areas where there's already water problems anyway. Putting in places like Arizona, Uruguay, places in Africa, all over the place. Like Arizona, all those different places where there's cheap land. But also scarcity in water where people are already having to be careful of how much water they use. So I think that is the major problem that we're seeing.

8:05 The amount of data centers that are being built in these areas that are already have water use problems, they're using more water. They need water for them, it's diverting water from other people. They're like people to actually like drinking water. And there's a huge problem there. So you have dry areas like Arizona, Spain, even Alberta, Canada, some places in Ontario, excessive groundwater pumping can cause salt water intrusion near the coast. So if you take the groundwater away, which pushes the salt water away from the coast along the coastline, salt water will come through into the groundwater where a lot of people get their drinking water from, right? So like coastal aquifers are, have problems. When sea water seeps inland contaminating the fresh water that can be bad for people's drinking water. So along those areas are bad. Some facilities in arid regions use desalinization, so like in the Middle East and so forth which creates brine discharges.

8:51 So really high salt water. So desalinization plants, they take in salt water. They take out the salt, and they bring in fresh water, but then they gotta get rid of the salt somewhere. So they gotta put it back into the ocean, which causes really high salinity in certain areas. So they'll have local areas that are affected on marine ecosystems near shore. There's a climate connection here. When you use different electrical sources. Right now a lot of the electrical sources are being used are coal, oil, and gas, which are fossil fuels, causing more CO2 to be in the atmosphere, warming oceans, causing acidification, and then stressing out corals, which you talked about a couple of episodes ago, where we're already past some of the tipping points. All the demand for energy and water is affecting the energy and water and the balance that is nature.

9:34 And because of the amount of use and the scalability of this, how much of a boom this has been, we are going to start to see impacts. The problem is nobody is measuring it. The reason why I'm talking about this in the first place is because I saw CBC, like the national news here in Canada talk about water usage and AI. I'm looking at a video. I watched this before, the BBC World Service Channel on YouTube. They talked about how thirsty AI is. I'm seeing a lot of different channels. I'm seeing a lot of different people on social media talk about water issues, which hasn't really been discussed in a large way because, you know, one, the big tech companies that are creating these AI tools that are helping us with productivity and helping companies and so forth, they don't wanna talk about it because it's an issue.

10:16 And we've seen this before in environmental fields where something goes big and it booms and then it grows so fast that people don't wanna talk about the bad stuff. Like when crypto came out, the amount of energy that these data centers for crypto to mine, the, cryptocurrency used up was really bad for the environment, actually caused more climate change. Like it caused more CO2 and more greenhouse gases to go into the atmosphere. Right. Now. Does it mean we can adapt to it? No, absolutely not. AI is not going anywhere. We are going to continue to use it in the future. If not, we are going to use it even more than we do now. We are just seeing the beginning of AI, and so there are certain things that we need to do and get done before all this gets too out of hand.

10:58 The first thing is we need to measure water. We need to measure how much water usage is happening at every single plant. That means regulations need to go into place. That's what's Gonna happen. We need to start using recycled and non-potable water. So non-drinking water for cooling. So Google and Microsoft are testing this, which is really great. We need to see liquid immersion or closed loop cooling. that cuts evaporation losses, because that's what happens too, when you bring water into these very high heated areas like these data centers, you're gonna get water evaporation so that water can't be returned to its original source. There's also geographical scheduling, so like where the location is is going to matter. And some of these AI companies, some of these big tech companies may have to pay more for the land that they do because there'll be more water available, right?

11:41 So like the location really counts. Location, location, location, it absolutely counts in this situation. Stop going to where the land is the cheapest because that's where the conflicts are really gonna happen. 'cause people are already suffering. That's why the land is cheap. ' cause there's probably not enough water. There's other problems that are at hand. Could be a poor area where people are struggling and we don't want to have these data centers causing these people to struggle even more just by getting clean water into their houses. Right. I think that's what it matters. That's another thing. Renewable energy, obviously integration to reduce indirect effects on water footprints is really helpful. Wind, solar, tidal, all these different types of renewable energy will help cut down any green house gas emissions. We're starting to level up and scale up on the amount of wind power, solar power, tidal power, renewable energy sources, which is great.

12:28 Of course they come with some of their own issues in terms of get buy-in from certain governments and so forth to really scale up and really help up. But we're starting to see increase in that as well for most governments. The big thing too is policy and accountability. So mandating public reporting of water usage effectiveness, which is WUE, integrating water stress maps into site approval processes. So making sure that when they put in where they wanna put the data centers, make sure there's enough water for everybody to use. That's a huge thing. Incentivizing any type of zero water or dry cooling innovations, they're coming. Those things are really happening. I'm gonna link to a video here the BBC World News. They actually talked about at the end of their video. And it's very, very new and still in research, but talking about putting some of the backup data centers in space because it's gonna be cool up there and they can put it in space.

13:15 We're already putting so many satellites in there. Whether that's good or not for the space, I don't know, but at least it takes some of the pressure off the ocean and off of different water sources around here. Let's be honest. Especially with fresh water, we are running out. It's not a dire need at this point, but later on in life, 50, 60, a hundred years down the road, there is gonna be a problem with fresh water, especially as water levels are going down. Evaporation is increasing due to climate change. We don't talk about how there could be a lack of fresh water in the future, but that's definitely a thing that we have to worry about. So ensuring that we can not only cool the planet to stop that from happening, but also making sure that these data centers don't cause extra stress on fresh water resources is going to be really important.

13:59 Now, you know, there's obviously connection to the ocean. Climate and hydro logic impacts ripple everywhere around the ocean. So having these data centers more inland, even if they're inland, they can still affect the ocean. Climate change is a huge thing about that. So indirectly. Water withdrawals can change river flows that feed estuaries, which can be a huge problem as we start to increase the amount of data centers that's in there or that are being built, energy related emissions accelerate ocean warming and sea level rise. Of course, desalinization, we talked about in industrial wastewater, create localized ocean salinity spikes When I was a private consultant, environmental consultant, worked for companies and we would do a lot of tests based on regulations that the government put out, provincial municipal, as well as federal governments.

14:44 When a sugar plant, let's just say they have discharged, you know, they got discharged, all their waste, and it goes through wastewater and it goes out into, say Lake Ontario or any of the Great Lakes. but it's heated water and it has a lot of nutrients in it. The big thing in the regulations that they have to make sure of an engineers are involved, water wastewater engineers are involved. They have to make sure that it acclimate very quickly and that there's certain numbers at the end of the pipe that matter, but the temperature difference coming out of those factories is unreal. It's definitely a difference. You see like this warm plume that comes out and you can measure it through temperature. We used to do like these maps, these temperature maps around the outlet for the outflow of this water and the sugar water basically. And you would see how hot it would get and you see all the fish come in because it's got a lot of nutrients and there's a lot of plants that grow there and stuff.

15:26 It's really, really interesting. But it has to be regulated. And the thing is it's not regulated. So the policy has to be put in place, and it's not rare and not surprising to see a lot of these tech companies be very cozy, cozy with different governments, not just the US but other governments because they're lobbying those governments so that they can produce as much AI and produce as many data sensors as possible, right? This is not a thing that hasn't happened in the past. It's not just a tech company that has provided these types of innovative lobbying kind of things. This has been happening in agriculture and everywhere. So I think that's an important thing. So the big thing is, is like what we can do about it, like the call to action for you is one advocate for transparency.

16:07 So communities should know how much water the local data centers use. First of all, do you know if you have a local data center? Do you know if there's a data center either being built or already built in your area, how much water is it being used? Ask your region, your county, wherever you are. Ask them why we don't have that information, why they don't have that information. Ask them to mandate that information. 'cause that's important to everybody. This is gonna affect everybody. Like I said, AI is not gonna go anywhere. Support cleaner, tech, you know, renewables, clean tech, recyclable non-potable water, but like support responsible AI companies that report their full environmental data. I don't even know what companies do and don't, I don't even know what companies build these data centers. If it's like Google or Meta or open ai.

16:50 Are they in charge of it or is it like a different company that does that, or different companies that do that. Educate others on the hidden water footprint. Share this video, share the information that you have. If you have your own information, if you know studies that have been done, share that. Streaming cloud storage, AI tools all depend on cooling. We're using it more and more, and this is gonna be huge. And then push companies to include water metrics in digital infrastructure permit. We probably didn't even know we had digital infrastructure permits, but they do, and we need to make sure that they are following the law. This is the big thing. AI can help us solve ocean problems. They can help us with productivity. If we build it and power it responsibly, we can really have a great impact on the world. AI and machine learning can help overfishing. It could help with identifying areas that are feeling the effects more identify faster ways to reduce climate impacts, all these types of things.

17:37 It could help us. We need to protect both the freshwater ecosystem and the oceans by designing the technology that respects the planetary limits. Like it really does. We are not above and beyond the planet limits. If the planet goes, we go and that's not good. And the goal is not to stop the AI. It's not to stop innovation. It's to make sure that innovation doesn't come at the ocean's expense and doesn't come at water expense. I think that's the big thing to do. This is not a video to say AI is evil and everything that's about it's evil. There's a lot of great things that can be done. Obviously there's a dark side to it as well, but there's a lot of great things that could be done in your own lives. As well as other people's lives and in the ocean's life. but we just can't let it go without being regulated for the environment, because if not, these companies will not take the ocean limits and take the health of the ocean and water to heart.

18:26 They just won't. They need to be regulated. So our governments need to get on this quickly and do it right now. Other industries that use a lot of water, the dominant one, agriculture. That uses over 70% of global fresh water. If you look at water usage for AI, it's like less than 2% at this point globally. Energy production is huge, about 15%, manufacturing and heavy industries, about 20% mining. There's a lot of water usage, food and beverage processing. There's a lot of water usage, construction and materials technology. Semiconductor manufacturing data centers, of course, is less than 1% of the total industrial water used globally, but growing fast. So we gotta put that in check, make sure that we're regulating it and make sure that it's going well.

19:06 and I think that's gonna be really important. So that's it for today's episode. I'd love to hear your thoughts on this. Let me know in the comments below because I think it's really important to start this conversation. if you're watching this on YouTube, comments below if you're listening to this on your favorite podcast app. Let me know how you feel. DM me on Instagram at How to Protect the Ocean. It's at How to Protect the Ocean, or Go To Speak up for blue.com/contact. Let me know how you feel. Just fill out the form, goes right to my email. I'll get back to you as soon as I can. I wanna thank you so much for joining me on today's episode of the How to Protect the Ocean Podcast. I'm your host Andrew Lewin. Have a great day. We'll talk to you next time in Happy Conservation.