Power Without a Playbook with Kristen Gonzalez

[00:00:04] Deep in western New York, between the cities of Buffalo and Rochester, maybe an hour's drive from each, you'll find the Tonawanda Seneca Nation. They are one of the six nations of the Haudenosaunee, the people of the Longhouse. They have lived on this land for generations, governed by a traditional council of chiefs and clan mothers. Their fight for this land spans the history of our country.

[00:00:30] In 1794, after the Revolutionary War, the United States signed the Treaty of Canandaigua with the Haudenosaunee. In it, the U.S. government promised not to disturb the Haudenosaunee nations, quote, for use and enjoyment of their lands.

[00:00:49] And then, in the 19th century, the Tonawanda Seneca Nation fought off removal and won back 7,549 acres of territory they still hold today. That land, its forests, its waters, its wildlife are central to how they live. It's the kind of place, the kind of history, that doesn't usually come up in conversations about artificial intelligence. But now, it's at the center of it.

[00:01:19] So I live here on the Tonawanda Seneca Nation. I live right in the area they call down below. That's Grandale Logan, a spokesperson for the tribe. Our territory is quiet. If you're not familiar with the area, our reservation kind of sits between this weird, like, tri-section of, I believe, Niagara County, Erie County, and Genesee County. Grandale's right.

[00:01:45] A drive through Genesee County, just next door, yields farmland, wetlands, and small towns. The county's population, about 58,000. But as quiet as this region is, there's a big, loud story unfolding here. In Genesee County, along the edge of the nation's territory, is a 1,250-acre patch of land called STAMP,

[00:02:12] which stands for the Science and Technology Advanced Manufacturing Park. For two decades, STAMP was backed by governors, senators, and tens of millions in public subsidies. To bring semiconductor manufacturing and green energy jobs to a part of upstate New York that had waited a long time for both. It didn't work. The STAMP site sat largely empty for years.

[00:02:38] The one major tenant that arrived ran out of money and stopped construction. So the county went looking for the next thing. That turned out to be data centers. It's existential for us. I mean, it butts right up against our territory, and there really is no major industry here. There is nothing as big as a data center. So Tonawanda Seneca Nation is paying close attention to this center.

[00:03:05] Because even though the characters may be different, they've heard this story before. And they know that their people, their ecosystem, will be dealing with whatever fallout comes from this project for years to come. So our issue is with the noise. And even our plant and animal relatives on our territory, we rely upon them, the animals, the plants, the trees. What kind of effect is it going to have on them down the line?

[00:03:30] As well as the light and increased traffic and any possible runoff or spills there at the data center site. It seems like the hydrology suggests that the way the water flows is towards our territory in the ground. And our people, we all have wells here in Tonawanda. So if there's any poisoning of the groundwater, well, that's going to eventually be a poisoning of our community as well.

[00:03:55] And according to the nation, that harm isn't just a neighborhood dispute. It's a violation of a federal treaty, the Treaty of Canandaigua. By having these environmental effects that bleed off into our nation is going to be a violation of this agreement. Because how can you put something that's going to have all these compounding effects next to our territory that's going to have impacts on our environment, which we use and rely upon?

[00:04:24] And how can you say that does not impact our free use and enjoyment? If we start to lose out on animals, on plants, on a lot of these things that we have in our environment, we begin to lose out on our traditions as well. There's a lot of things that will atrophy because there's going to be practices we won't be able to do if this data center goes through. Stream Data Centers, a Texas-based firm, wants to build a campus at Stamp. And what's happening here is not an accident.

[00:04:55] Unlike the data centers that power cloud functions, which need to be close to cities, AI-specific data centers can be built anywhere. So developers go where land is cheap, flat, and close to power. And rural communities across the country are waking up to find themselves next in line. Still, these communities are pushing back, including the Tonawanda Seneca Nation. They sue to stop Stream's development.

[00:05:24] Soon after, the county rescinded its approvals. So we celebrated that for a time, of course, because we have really done a lot of good work here at the Stamp site using the power of delay. And so by delaying their process further, personally, and like what the belief is for our nation is that does a lot. That does a lot in trying to convince Stream Data Centers to not come to the Stamp site.

[00:05:53] Then the developer came back, bigger. So it was kind of a sense of like, well, here we go again. You know, like we did it the first time and we figured, well, we'll do it again. And so there's a lot of positivity coming along with us in everything that we've done up to this point. And so the fight needed to get bigger, too. This is TED Tech, a podcast from TED. I'm your host, Sherelle Dorsey.

[00:06:22] This is part three in her series on data centers. We've heard the case for data centers as a civil rights issue and the toll they're already taking on water. You can find previous episodes on our feed. Today, we want to explore accountability. What does it actually look like to fight back? What does it look like for a data center to move in responsibly?

[00:06:45] The Tonawanda-Seneca Nation's fight is one front in a much larger battle, one that's now moving through Albany, the capital of New York. The mantle has been picked up by State Senator Kristen Gonzalez, the chair of New York State's Internet and Technology Committee. She's been watching what's happening in counties like Genesee, and she's trying to build a regulatory framework fast enough to matter.

[00:07:11] Earlier this month, New York state lawmakers passed a bill that would pause data center construction like the one planned in Genesee County. If the governor signs the bill, New York will be the first state to do this. One year, no new permits, time for the state to figure out what it's actually been agreeing to. For today's episode, I spoke with Senator Gonzalez, who was behind the bill, about what it actually looks like to regulate an industry that moves faster than the law,

[00:07:41] and what accountability means when you're up against the most powerful companies in the world. That conversation is coming up right after a short break. Senator Gonzalez, you've been working on data center legislation for years. So what finally made this the moment it moved? I think this is a result of years of organizing by a broader environmental justice coalition.

[00:08:11] We introduced one of our first bills on data centers actually a couple of sessions ago. And that was because we saw the early warning signs that hyperscale data centers were showing in other states and recognized that in an affordability crisis and a climate crisis, it was important to make sure that any costs associated were being shouldered by the companies that were actually building out these data centers and not by everyday New Yorkers and certainly not by my own constituents who are struggling with some of the highest utility bills they've ever had.

[00:08:39] And what has also become increasingly clear is the number of projects and the potential added energy strain on the grid is on a level with which we do not have an actual plan to meet. So it is incumbent on us as regulators to pause development, to have a clear plan, and to also while we are doing this type of planning for a positive energy future,

[00:09:06] one that actually puts New Yorkers first and our best interests first, that we also are creating the infrastructure and processes to deal with these hyperscalers. What I'm coming to understand is that all data centers are not created equal. Can you just provide some understanding of what a hyperscale data center is and how that might differ from other data centers? As defined by this bill, a hyperscale center is 20 megawatts and above.

[00:09:34] And, you know, as someone who didn't have a background in energy, I didn't really know how big a megawatt was. But to contextualize it, it could, one, a single megawatt power up to a thousand homes. And so we're talking about not only an incredibly energy-intensive facility, unlike the data centers, many of them that are online today, but we're also talking about one that is massive in size.

[00:09:59] What we're seeing in other states are data centers, or hyperscale data centers, rather, that are the size of several football fields. So it's really important for people to understand that the reason we're trying to meet this moment is in order to power generative AI, we're seeing a type of data center that we have not had to necessarily accommodate on our grid or in our communities before. So that's what takes the immediate action from our state and local governments.

[00:10:28] So what does it actually look like to hold that line in a legislator when big tech is in the room and has had so much leeway, has kind of bypassed through various loopholes, right, to kind of super speed and hyperscale these data centers in an unprecedented way? Yeah, I think you bring up a good point.

[00:10:52] Big tech, and I say big tech, especially in this context, because the companies that have the resources to build out these massive facilities are the biggest of the big. They are your Metas, your Amazons, your Googles, these tech goliaths that have an almost infinite amount of resources. And they are so used to being able to be in the room when regulation is being considered and either prevent it from getting passed or write the rules for themselves.

[00:11:18] So this is a really significant moment for New York State because we're taking some of that power back and leveraging government as it should be used to actually put people first above their interests. But it's something I've certainly seen, not just particular to data centers, but across the spectrum of AI regulation. It's not enough to just regulate the build out of data centers. Again, these massive centers are not only for AI, they're actually for generative AI or agentic AI, the use cases of which are still being worked out.

[00:11:47] And I personally do not believe should be considered inevitable and should be regulated to protect people. Absolutely. And I want to go into who's actually being targeted for some of these data center developments. We're starting to see somewhat of a pattern of where these are being built and developed. But what's the pattern that you're seeing across New York? Like who's being targeted for development and why those communities specifically?

[00:12:13] So I don't think there is a ton of transparency into where every project is being cited or where there's additional speculation or interest. Now, oftentimes these projects are, because they are incredibly large, they're going to be in more rural counties. They're going to be in upstate New York. They're also likely going to be cited near large bodies of water because water is needed for things like cooling.

[00:12:39] So we already kind of see, to your point, that they are cited next to indigenous protected communities. And we need to be, again, clear that that is not acceptable. But as I think we've been investigating across the country, we've noticed that increasingly they are cited next to communities of color, lower income communities. And that's why we wanted a state bill that empowered local governments and kind of even the playing field if they want to negotiate and potentially bring on a hyperscaler nearby.

[00:13:07] Now, some of the arguments here for adding data centers, particularly in counties, some officials would say, well, they were going to bring jobs, right? There's going to be some level of tax revenue. But when you look at the actual record, is that argument holding up? This idea of bringing jobs is one that we hear a lot. And to your point, I think we need to think a little bit more about what these jobs are and what they mean to know if that's a true statement.

[00:13:34] It is true that these hyperscale projects bring thousands of construction jobs. But to be clear, these are short-term jobs. When it comes to long-term jobs, so being able to, once they're operational, actually provide new job opportunities for local communities, those tend to be very limited and very few.

[00:13:56] And if you compare the amount of money that's invested in the project to the number of long-term jobs that are created, you're oftentimes talking about tens of millions of dollars per single job. I want to dive here into the moratorium to specifically understand what this data center bill means, what you're proposing, what you're hoping to accomplish with it. One year feels like such a short period of time to try to get a lot done around this idea of responsible data center development.

[00:14:23] What has to happen in that one-year horizon for everything that you are hoping to accomplish to be accomplished? You know, this one year is reflective of a balance of interest. We had a three-year moratorium in our initial bill. We wanted to make sure—I think it's important to note that I personally believe we need more time than one year. And we want to make sure that we are not completely disincentivizing investment in this technology in New York.

[00:14:51] It means that we at least have the next 12 months to do real community engagement. Part of this bill is the environmental impact report that will allow us to start collecting data on the impacts of data centers across the state, but then, in addition, have public hearing sessions. That is not something that has been done so far. And I know New Yorkers have a lot of opinions about where data centers should be located, when they should be used, how they should be used.

[00:15:19] And I'm hoping in those public engagement sections we'll actually get a lot more information about what a policy future should look like so that at the end of these 12 months, when we're back in our legislative session, we can introduce new bills to regulate hyperscale data centers. The opposition saying that once you place this moratorium, it tells people that New York is closed for business.

[00:15:43] And so, you know, what are you saying to the officials who are believing that this signals some level of a volatile business environment for those who are looking to build or expand in New York? Yeah, to be clear, we're not imposing a ban on data centers. We are pausing the permitting from the Department of Environmental Conservation of hyperscale centers that are 20 megawatts and above. So I want to be clear, we're not getting in the way of innovation.

[00:16:12] We're not banning any technologies or banning data centers. But we are understanding the unique challenges that hyperscalers pose. I would also say that, you know, a lot of my colleagues have agreed that one-year pause is not a signal necessarily that New York is closed for business. And I think I mentioned, you know, initially that three-year moratorium was negotiated into a one-year moratorium, which shows that we're willing to meet some of the folks who are worried about industry not investing in the future, meet them halfway at least.

[00:16:41] And even, I think, generously at one year and, you know, balance again, serving our constituents well, while also making sure that in the future there's development. The last thing I'd say is, as someone who worked in technology and has also worked in the private sector, the idea that data centers will just go to other states isn't necessarily true.

[00:17:05] It is beneficial for these large tech companies to build hyperscalers near large consumer bases because it helps with things like latency or how quickly their products are working or, you know, their large language models are giving you a response. So as the consumer base in New York continues to grow, their interest in building a hyperscaler to be the fastest and best product in the market will only increase.

[00:17:34] So I see that as being true now, but even truer in a year. And I do not think we'll see every hyperscaler project leave the state. So Senator Gonzalez, if this bill gets passed, if it gets signed into law, it will be the first in the nation. I think we mentioned that earlier in the conversation. What do you want it to prove is possible in this country?

[00:17:57] I'm really glad you asked that because in a lot of these conversations on the omnibus bill and the moratorium, I've been trying to intentionally take a step back to tell a broader story about what it means for power for everyday people.

[00:18:14] Right now, these large AI companies, these tech goliaths, are not only lobbying state legislatures, but they're lobbying the federal government to completely preempt states from legislating. They're also lobbying the federal government to make sure that they are either under-regulated or are able to not meaningfully be challenged in the ways that they operate today.

[00:18:37] I am also really disturbed by the amount of money they are now not only putting into lobbying, but putting into races. They have spent millions of dollars to either help prop up candidates that they like or to run negative campaigns against candidates that want to regulate them.

[00:18:58] So their influence, not only in our regulatory processes, but in our democracy, has become incredibly outsized. So this fight isn't just about a positive energy future for New York, isn't just about regulating data centers.

[00:19:14] It really is about establishing that New Yorkers should be the ones in the driver's seat, everyday people should be the ones in the driver's seat, and that collectively we should be setting the rules of the road, so to speak, for the future of tech regulation, for data centers, for AI safety, and more. What has this been like for you, regulating in real time? You know, technology is constantly developing. It's changing faster than most of us can keep up with. But how can local leaders keep up?

[00:19:45] Well, I think government works best when levels of government are working together. I think local leaders deserve a state legislature that are passing bills that will empower them in the negotiation process. Local leaders have signaled that they want to see the state act in this area, but also, of course, want to continue to have the local power to decide whether or not permits are given.

[00:20:12] So this bill does not, for example, take away any local government's ability to issue their permits as they would normally issue permits. But that community benefits agreement piece will make sure that they are able to maximize any benefits for their own constituents moving forward. That is how we see our state legislature being supportive of our local governments.

[00:20:32] I think local leaders, though, are also really aware of the strains that hyperscalers pose on their communities and have heard from constituents that they do not want to see these hyperscalers.

[00:20:46] So a lot of the conversation hasn't only just been around regulation, but it's also been about organizing and the future of creating these coalitions that are incredibly diverse, that extend party lines, extend demographics like age groups, ethnicity, extend even economic classes,

[00:21:04] and really creates a unifying issue by which these local leaders can not only engage their community, but then build power in their community in the face of some of the wealthiest lobbies in the country. Very well said, Senator Gonzalez. I just want to thank you for taking the time to join me here on TED Tech. Thank you so much for having me. Senator Gonzalez is fighting this battle in the room where the laws get written.

[00:21:32] But there's another version of this fight, one that happens before a single permit is filed, before a shovel ever hits the ground. We have the last piece of today's episode, a powerful talk about curbing the costs of AI coming up after this break. AI sustainability expert Sasha Luccioni studies the environmental cost of artificial intelligence for a living.

[00:22:02] Here's what she found. Most of what we're told is inevitable by AI's energy use isn't inevitable at all. It's a choice. A choice made by companies that have so far had very little reason to choose differently. Here's Sasha Luccioni on the TED stage making the case for what it would actually take to build AI we don't have to apologize for. Revolutionizing science, turbocharging productivity, even solving climate change.

[00:22:30] AI has been promised to transform the future of humanity. Or it's set to bring about the end of humanity as we know it. It really depends on who you ask. In my opinion, both of these statements are wrong. And what they do is they distract us from the real issue at hand. We're doing AI wrong at the expense of people and the planet.

[00:22:52] As it stands, a handful of large corporations are using huge capital to sell us large language models, or LLMs, as the solution to all of our problems. Possibly because they think that they'll bring about superintelligence, emotional intelligence, basically whatever flavor of intelligence is trending in Silicon Valley these days. And in this race, they're building more and bigger data centers, the people and the planet be damned.

[00:23:21] Meta is set to build a data center the size of Manhattan in the next few years. Part of an investment of hundreds of billions of dollars towards a quest to develop superintelligence. OpenAI recently announced the first phase of their Stargate data center in Texas. Once operational, it's set to emit 3.7 million tons of CO2 equivalents per year, as much as the whole country of Iceland.

[00:23:49] XAI is currently being sued by the residents of South Memphis because of the air pollution caused by their 35 questionably legal gas turbines, which are powering its data center, Colossus, exacerbating the health issues of the city's most vulnerable residents. And yet, for years, activists and scientists like myself have been sounding the alarm when it comes to AI's increasing unsustainability. Does this ring a bell? Remember Big Oil?

[00:24:18] Well, now we have Big AI following the exact same playbook, using more and more resources, building bigger and bigger data structures, and selling us the narrative that this is somehow inevitable. But what if we could learn from the lessons of the past and use them to build a future in which AI is giving back to the planet instead of taking away from it? A future in which AI models are small but mighty, in which they are both better performing and more sustainable.

[00:24:45] To do this, we have to take back the power, pun intended, from the big AI companies and put it back into the hands of the developers, regulators, and users of AI. Today, we use AI as if we were turning on all of the lights of a stadium just to find a pair of keys. Using huge AI models, trained using the energy demands of a small city just to tell us knock-knock jokes or help us figure out what to make for dinner.

[00:25:13] This is driven by a bigger is better mentality. This has become somewhat of a mantra in AI. Bigger models, more compute, bigger data sets, more energy equals better performance. And the pinnacle of this approach are LLMs, models like chat GPT, which are trained specifically to be general purpose, able to answer any question, generate any haiku, and act as your therapist while they're at it.

[00:25:37] But this performance comes at a cost because models that are trained to do all tasks use more energy each time than models that can do a certain task at a time. In a recent study I led, we looked at using LLMs to answer simple questions like, what's the capital of Canada? And we found that compared to a smaller task-specific model, they use up to 30 times more energy. And as this energy use grows, so does their cost.

[00:26:06] Essentially, with the number of organizations that can afford to build and deploy what's considered state-of-the-art AI shrinking, becoming limited to a handful of big tech companies with millions of dollars to burn, while startups, academics, and non-profits are all left in the dust. So now this handful of big AI companies, largely gathered by the move-fast and break-things mentality, decides the future of a technology that can impact the lives of billions of people.

[00:26:35] But in the background of all this hubbub around the DeepSeqs and the ChachipiTs of the world, a revolution has been quietly building in recent months. This revolution is driven by small LLMs, which are also language models, but that are orders of magnitude smaller than traditional LLMs. The smallest of this family has around 135 million parameters, making it 5,000 times smaller than DeepSeq's model. These models are flipping the script on the bigger-is-better mentality

[00:27:04] by using less data, less compute, less energy, and still having the same level of performance. The data used to train HuggingFace's small LLM models was carefully curated to be 60% educational web pages, explicitly chosen based on the quality of their content. This also means that the models that are trained on this data are less likely to produce misinformation or toxicity when we query them. And since the models are so small,

[00:27:31] they can run literally on your phone or in your web browser, giving you access to state-of-the-art AI in the palm of your hand without needing massive data centers. And above and beyond environmental impacts, they also have benefits when it comes to cybersecurity, when it comes to data privacy and sovereignty, giving users more power over the AI that they're using. And since they're smaller and cheaper to train, they give smaller AI companies the ability to connect with a community

[00:28:00] and to compete with big AI companies because they can actually afford to be training and deploying these models and adapting them to different uses and then sharing them back with the community, proving that reduce, reuse, recycle also applies to AI. But the truth of the matter is that there's more to AI than just small LMs. And if we really want to make AI more sustainable, we have to be thinking beyond LLMs to using all sorts of different approaches that can be really useful in our fight against climate change.

[00:28:29] Because, sure, Chad GPT can tell you which countries signed the Paris Agreement, but it can't predict extreme weather events, which requires an understanding of the physics of weather patterns and geography. And sure, Claude can explain the whys and the hows of climate change, but it can't help a farmer decide when to plant their crops based on temperature, humidity, and historical weather patterns. There are so many other approaches in AI that use less energy and still are really useful in our fight against climate change.

[00:28:58] For example, recently, a team of researchers funded by NASA trained the Galileo models, which can be used for all sorts of different tasks, from crop mapping to flood detection, without needing specialized hardware. This makes them accessible to governments and nonprofits. And Rainforest Connection uses AI to do bioacoustic monitoring. That means that they listen to the sounds of rainforests across the world, identify species, and even detect the sounds of illegal logging in real time. Their AI models are so small,

[00:29:27] they run on old cell phones powered with solar panels. And Open Climate Fix uses AI to analyze satellite imagery, weather forecasts, and topography data to predict the output of solar and wind installations, allowing us to move forward to decarbonizing energy grids around the world. This includes data centers, because currently they're powered by mostly coal and gas, but they could be renewable if we had the right tools. But another problem is, as users of AI,

[00:29:56] we don't know how much energy an AI model is using or how much carbon it's emitting when we use it. That means that we can't make decisions with sustainability in mind, as we do for the food that we eat or for how we get around town. This led me to create the AI Energy Score Project, in which we tested over 100 open-source AI models across a variety of different tasks, from text generation to images, and we assigned them scores from one to five stars based on energy efficiency.

[00:30:25] So say that you forgot the capital of Canada again. It's Ottawa. You could use a model like Small Alarm, which would use 0.007 watt-hours to give you that answer. Or you could use a model like DeepSeek, which would use 150 times more energy for that answer. But sadly, big AI companies didn't want to play ball and evaluate their models with our methodology. And honestly, I can't blame them, because the truth might only make them look bad.

[00:30:56] Because currently, we don't have the laws or incentives that we need to encourage AI companies to evaluate the environmental impacts of their models or to take accountability for them. The EU AI Act started this process by introducing voluntary disclosures around the energy and resource use of AI models. But enforcing this act in Europe and eventually writing laws like this across the world will take time that we simply don't have given the speed and the scale of the climate crisis.

[00:31:25] But the good news is that we don't need to stay hooked on the AI sold to us by big AI companies today as we've stayed hooked on the coal and plastic and fossil fuels that have been sold to us by big oil for all these decades. And in fact, instead of believing that the future of AI is already written, that it consists of huge LLMs powered by infinite amounts of energy that will somehow result in superhuman intelligence

[00:31:54] and magically solve all of our problems, we can take back the wheel and shape an alternative future for AI together. A future where AI models are small but mighty, where they run on our cell phones and do the task they're meant to do without needing huge data centers. A future in which we have the information we need to choose one AI model over the other based on its carbon footprint. A future in which legislation exists that makes big AI companies take accountability

[00:32:23] for the damage that they're causing to people and the environment. A future in which AI serves all of humanity and not just a handful of for-profit tech companies. With every prompt, every click, and every query, we can reinvent the future of AI to be more sustainable together. Thank you. That was Sasha Luchoni at TED Countdown.

[00:32:51] So what does accountability with teeth actually look like? It looks like a nation that draws from history and fights for their community. It looks like a local politician winning a year of public hearings for their constituents so they have the space to learn and choose their own futures. And it looks like a researcher who knows the stakes and has a solution. None of that stops the data centers. Not yet. But it changes who's in the room when the next one gets built,

[00:33:20] and it changes who has to answer for it. That's our show. We'll be back next week with another episode in our series on data centers and the infrastructure behind AI. And go check out the other episodes in this series if you haven't already. You can find them all on the TED Tech feed. TED Tech is a podcast from TED.

[00:33:49] This episode was produced by Rahima Nasa. Our editor is Alejandra Salazar. And the show is fact-checked by Julia Dickerson. Special thanks to Constanza, Gallardo, Daniela, Bellareso, Maria Ladias, Tanzika Sangmanivan, and Roxanne Heilash. If you're enjoying the show, make sure to subscribe and leave us a review so other people can find us too. I'm Sherelle Dorsey. Let's keep digging into the future. Join me next week for more.