Subscribe
Share
Share
Embed
This episode pulls back the curtain on the AI gold rush, the data centers, the water bills, the NDAs, and what it all means for rural communities that rarely have a seat at the table.
AI sounds like the future, but the costs are landing unevenly, especially outside major cities. Heidi and Joel join Dr. Emily Bender and Dr. Alex Hanna to dig into the real, and rarely discussed, toll of our digital infrastructure boom, from secretive corporate deals to environmental strain, and ask the question nobody in Silicon Valley wants answered: who actually pays the price?
In this episode:
- The gap between AI hype and reality, and why it matters
- Data centers sprouting faster than the regulations meant to govern them
- The true costs to energy, water, and local infrastructure that corporations aren't advertising
- Public resistance, NDA nightmares, and the political pressure to build fast
- Why regulation hasn't kept pace and how communities are pushing back
- The risks of AI overreliance, hallucinations, and why source-checking matters
- Where international regulation stands and the gap in U.S. policy
Guests:
- Emily Bender - Twitter | University Profile
- Alex Hanna’s Website
- DAIR Institute
The AI boom isn't slowing down, but neither are the people asking the hard questions. Tune in, get informed, and maybe think twice before you trust the hype.
The Hot Dish is brought to you by the One Country Project. To learn more, visit OneCountryProject.org, or find us on Substack (Onecountryproject.substack.com), and on YouTube, Bluesky, and Facebook (@onecountryproject).
Creators and Guests
Host
Heidi Heitkamp
U.S. Senator Heidi Heitkamp served as the first female senator elected from North Dakota from 2013 – 2019. he is the founder and Chair of the One Country Project, an organization focused on addressing the needs and concerns of rural America. Heidi was recently named the Director of the Institute of Politics at the University of Chicago, a university she has long been committed to and a place where she enjoys engaging with students over civic discussions while encouraging them to seek opportunities in public service to our country. Heidi also serves as a contributor to both CNBC and ABC News.
Host
LA
Producer
Laura Anderson
What is The Hot Dish?
Former U.S. Senator Heidi Heitkamp and her brother, KFGO radio talk show host Joel Heitkamp, engage in animated discussions with newsmakers, elected leaders, and policymakers who are creating new opportunities for rural Americans and finding practical solutions to their challenges. Punctuated with entertaining conversations and a healthy dose of sibling rivalry, The Hot Dish, from the One Country Project, is informative, enlightening, and downright fun.
Heidi (00:04)
Welcome back to the hot dish comfort food for rural America. I'm Heidi Heitkamp.
Joel (00:09)
And I'm Joel Hickey.
Heidi (00:11)
AI is a hot topic right now. Everybody's talking about it. You see graduates, even the vice president getting booed ⁓ when they talk about AI in the future because there's a lot of uncertainty of what that looks like going forward. So I think it's really important, especially what's happening in rural America with the building of data centers and controversy surrounding that. A lot of jurisdictions, in fact, last count about 70 banning any form of ⁓
building ⁓ AI centers or data centers. And this idea that we desperately need ⁓ this kind of computing power for the economy of the future, ⁓ it doesn't seem like rural America is getting the message. And so we decided today that we would invite a couple experts to our podcast, Emily Bender and Alex Hanna. They're the authors of a recent book, The AI Con.
how to fight big techs hype and create the future we want. Alex and Emily, thank you so much for joining us on the hot dish to talk about this important issue, which has gotten even more important in rural America.
Alex Hanna (01:23)
Thanks for having us.
Emily (01:24)
Real pleasure to be here.
Heidi (01:26)
Well, Joel, you have spent the most amount of time talking to folks, lots of controversy in North Dakota, but North Dakota's not alone. So maybe you can kick it off and give us a rundown on what's happening in our state, but also reflects what's actually happening across the country.
Joel (01:45)
Well, I'm going to start with Dr. Hannah just because I know the answer is going to be the same from both. But I think one of the things that people struggle with the most is understanding what AI is. And so, Dr. Hannah, give them the short version.
Alex Hanna (02:01)
So AI doesn't refer to any singular technology. AI is actually a marketing term. It's actually quite an old marketing term. And it wasn't even the marketing term of choice until about 10, 15 years ago. And so when people talk about AI, it can refer to a bunch of different things. It could refer to things as basic as a classification algorithm that tells you the difference between whether there's a cat or a dog in the picture.
or it could refer to a recommendation algorithm, like what to watch next on YouTube. But I think a lot of people conflate AI with ⁓ chatbots. called, you know, and this type of chatbot that we have now is called the large language model. That is the type of model that is behind chat GPT and Claude and Copilot, all this stuff that you get see shoved into all these different products that we're seeing nowadays. So that is
what that seems like it is. And then we have this particular thing, large language model, which is a very souped up next word prediction machine. And I'm going to pass it to Dr. Emily Bender, my co-author, to explain how those work a little bit.
Emily (03:17)
Yeah, so part of the reason everyone's so excited about so-called AI right now is that we have systems that seem to speak our language. The large language models are basically systems for mimicking the way people use language in lots of different contexts. But it only makes sense because we make sense of it. And when Dr. Hannah says that this is about souped up autocomplete, you can remember that game that we used to play with the smartphones where you type in, I'm sorry, I'm late, I, and then just pick the middle word over and over again.
And everyone got slightly different answers because what was going on was that your phone was tracking the frequencies with which you use different words in different contexts. And that was a little local language model on your own phone, which is why you got a different answer from your friend. That's all that's happening here, except that it's trained on absolutely enormous amounts of text. And so it's much harder to track, it's saying that word because it's trained on this text, because we actually don't even know what's in that training data.
Heidi (04:01)
for this treatment. Absolutely.
Well, it's incredible and, you know, true disclosure, I use it all the time. I mean, ⁓ it's very helpful as a first draft. It's very helpful to kind of help you ⁓ kind of outline your thoughts. It can be incredibly valuable as a tool that goes beyond your typical search engine, but
There's a whole lot of people who look at this and say, it's going to take over the world. So either one of you can ⁓ respond to, know, what, what, what, then the Terminator, what was it? It was Netscape or what, was the Skynet, Skynet. Yeah. That, that Skynet is coming.
Emily (04:51)
You know, Skynet was the villain. Yes. Yeah.
Alex Hanna (04:53)
Although, although, although, although saying Netscape is not
as far as you think, but for different reasons.
Joel (04:59)
you
Emily (05:00)
Yeah.
Heidi (05:00)
Yeah, I know. Maybe that was a Freudian slip, you know, Skynet. So we're to Skynet, you guys.
Emily (05:07)
No, absolutely not. There's a lot of danger in this moment, but it has to do with the actions of people and not at all with these enormous conglomerations of predict the next word, because predict the next word is just what's the likely next word that is not sentient. It's not conscious. It doesn't have any desires of its own. It just makes paper mache out of the training data.
Alex Hanna (05:29)
Yeah, and I think there's a lot of ways in which there's a lot of media coverage that's suggesting that there is kind of sentience. There was a New York Times article that came out yesterday, said something like, the title was like, the first quote unquote AI actor wants you to know that she's tired of all the criticism or what she wants you to know. And there's so much that's wrong with that. First is that it is an algorithm. It doesn't want anything.
Heidi (05:49)
Ha ha ha.
Alex Hanna (05:58)
And then also this reference to it with these pronouns, like she and her and he and whatever. I mean, these are technologies and tools and there's lots of issues with them, but there's no risk of this kind of takeover that we see. That is purely in the realm of science fiction.
Joel (06:16)
If you think they don't know what they're talking about here, just understand their co-authors of the AI con, how to fight big techs hype and create the future we want. And they do their own podcast as well called Mystery AI Hype Theater 3000. So there I made sure to get that out there because you guys own it. You deserve it. ⁓ I want to throw this out to you because as you described AI.
Many of the people that listen to the hot dish from the rural area think of ⁓ where it's stored and not the product itself. And so, you know, we're being, ⁓ for lack of better way of putting it, we're being somewhat underappreciated when it comes to the storage of it. And I think it's ahead of ⁓ the understanding of the industry itself.
⁓ For me, ⁓ I want to say, Dr. Bender, that we end up kind of carrying the water out here in the rural area. And I'm curious what you think about that.
Heidi (07:21)
So to speak, yeah.
Emily (07:22)
Yeah.
I mean, my impression of where the data centers are being situated and just backing up a second, the illusion of progress that we are seeing from these companies has to do with making ever larger models. That's their only trick. Let's make this bigger and bigger, but bigger and bigger means more data, more compute. You need the big data centers to store that and do all that processing. And those big data centers require a lot of energy and they require a lot of water for cooling, as you say. And the pattern that I'm seeing
is not that they're being cited in places where energy and water are abundant, but where they are under regulated. And that is, I think, really unfortunate and probably describes ⁓ much of rural America, not least because you have, people are more spread out, right? It's harder to mount resistance. It's harder to have consistent regulation. If this county doesn't want it, we're going to go one county over. But the water tables aren't delimited by counties, right?
Alex Hanna (08:19)
Yeah. And I would just add onto that too, in terms of what we're seeing in other parts of rural America. So mother lives in a pretty rural ⁓ area of Ohio and the county that went solidly for Trump and has for the past few elections. And there's actually a township in that county past the first ⁓ data center moratorium in the state of Ohio. And that was surprising. Now that there's other counties
other townships in that county also looking at moratoria. And so it's a consistent issue and there's a lot of outrage that comes with the water, the noise, the air pollution, and then the carbon cost. And the conversations are looking different in different communities, but they're all kind of wrapped around this body of different concerns. another thing about Ohio, I had a state senator in Ohio that was telling me, Ohio ranked seventh in data center construction.
⁓ but it is 27th in terms of the tax receipts. And that's because Ohio currently has a sales tax exemption on data center construction. And so we're seeing the political response look different in different places. ⁓ Some of the weaker policy interventions include ⁓ things like doing an exploratory study, which is important. Yes, that is a first step, but more of the active
⁓ interventions are having a moratoria, understanding if there's a way, if there's building, there's a lot of agreements and benefits that do need to come and accrue to the municipalities and the counties and the states prior to these projects going forward.
Heidi (10:01)
Yeah, but one of the big problems is that the political leaders of these organizations, not the people. ⁓ In fact, there's places where the people have literally recalled political leaders who have allowed this to go forward. ⁓ The only real advantage you have, because there isn't employment after the construction employment, the only advantage you have is you're building a tax base, right? I mean, there's a lot of investment that...
theoretically, if you put it in a county, should wipe out the rest of the county's property taxes, should be huge in terms of sales taxes on the equipment that goes into a data center. But it's all being given away. And it's being given away at the same time that they're requiring these same political subdivisions, in my opinion, as a former attorney general of the state of North Dakota, who enforced the open records law. In violation of open records law, they're forcing NDAs.
non-disclosure agreements, which are highly, highly controversial because you can't know what they've given away. I mean, think about it. You can't know what they're going to do. You can't know how much water they're going to use or how much electricity and what that means in terms of stress on the grid because these same idiot political leaders have basically waived the public's right to know. so, you know, how, how, as you look at this, I mean, let's put aside the
Alex Hanna (11:20)
Yeah.
Heidi (11:26)
the value of AI and what that means. Because when you look at public policy on AI, I always say it is hurry up, hurry up, hurry up, and that's being pressured because of China. China's gonna beat us, China's gonna beat us, and then you get the resistance, stop, stop, stop. And so we don't have consistent national or statewide policy. But these NDAs, going back to the data centers, these NDAs are very troublesome, I think, for anyone who believes in a transparent government.
Emily (11:56)
100 % and you often don't even know who's contracting like who's the data center for who's actually paying to have it built? That's another thing that's hidden and I think that that is hidden partially because these companies don't have a good reputation and they want to avoid resistance coming from the people but you are absolutely right that People deserve information about what's gonna happen with their resources that they need right? we need water to drink and grow food and to live and you know, that is
Alex Hanna (11:56)
Yeah.
Emily (12:24)
higher priority, it ought to be higher priority than whatever these data centers are. But if the people don't have information about what's going in, it makes it much harder to fight. So I think you're absolutely right that fighting for transparency is a really important first step.
Alex Hanna (12:36)
centers have been, I mean, we've had data centers in the past. mean, for the internet to run, you need some kind of servers and those servers need to live somewhere. We use this metaphor of putting things in the cloud, but.
The cloud has to live somewhere, as Jill said. And so ⁓ those data centers have been much smaller. ⁓ Hyperscale data centers is this terminology, which it doesn't have an actual number associated with it. But there are these data centers just of huge size, which are much larger, which ⁓ are energy intensive, water intensive,
Heidi (12:51)
you
Alex Hanna (13:13)
a huge ⁓ ecological and physical footprint that get located ⁓ and are used to build that internet infrastructure. So prior to the data center boom, this is when Google was getting much larger. They wanted to build one of these hyper-scale data centers in the Dulles, Oregon. And they had one of these NDAs where they actually had an agreement with ⁓ the local government to prohibit even
stating how much water they would use. And it wasn't until the Oregonian had sued the city government that they actually found out how much water consumption would be used by that data center. It was something like half of what the city was using at the time. And so you're seeing this as a strategy that's being used more and more. And also these companies are not forthcoming about how much energy their huge models are taking.
They don't say there's not a water report. There's not an energy report. There's not a report on how much ⁓ small air particulates will be output at a data center because of backup diesel generators. ⁓ There's all these layers of transparency between the NDA, between the model outputs. And it's just this handshake deal, whatever, between whoever is allowing these to go forward and the companies.
Joel (14:40)
Well, let me just say this in case you're wondering if you're right, which you know you are. You're right. And let me give you an example. A little town called Ellendale. It's in my backyard, not too far away. ⁓ Ellendale struggled. They struggled for jobs. They struggled for economic development. The number of kids in the high school or school system kept going down and down and down. And I think it still is. But their property tax was their main concern.
Heidi (14:45)
Yeah.
Joel (15:07)
And so what did they do? They went in and they got a bunch of non-disclosure agreements signed from the local officials. But the one thing that we're not talking about here and you mentioned, you know, the power grid, you mentioned the water. All of those things are accurate. The only thing we haven't talked about is the roads. You know, it's usually in township areas and the township roads are just getting wiped out. But I mean, here's the thing. They prey upon the fact that these small areas have not
had economic development and the political leaders have been desperate for economic development and then they get it or at least they get what they hope to get. The biggest thing to me that happens after they strike and I want to see if this is true nationwide as you two look at this from a bigger broader picture is how fast it happens. Once it's announced. I mean you can have six seven cranes there within a week or two.
I mean, there's no going back. There's no protesting. There's no town hall meetings. There's no taking. And you mentioned county to county. We had a small city ⁓ in in Harwood, North Dakota, ⁓ which was in trouble. The mayor was for signing a non-disclosure agreement and having this big project right in the backyard when they heard bad things about it. And then the city of Fargo say, no, we want it. And so pretty soon these two communities were fighting each other over it and
Heidi (16:28)
and so pretty soon these two reveals are fighting each other over it
and you know instead of being ticked about it they're fighting over who gets it and so I'm
Joel (16:35)
you know, instead of being ticked about it, they're fighting over who gets it. And so I'm curious
what your response to that is.
Emily (16:44)
It seems to me that part of the appeal is that this is supposed to be high tech, big tech, right? And if you're thinking about looking at what's going on in say California and it seems like the tech industry is driving so much economic growth, why wouldn't you want some of that in your town? And the problem is, as you've been saying so clearly, that's actually not what it is, right? There's not jobs past the construction jobs. And my understanding is that many of those construction jobs aren't local either. They bring people in to do the construction work. And then if you've waived the sales tax,
Heidi (17:04)
my understanding is that many of those construction jobs are on the floor by them. They bring people in, do the construction work, and then if you waive the sales tax,
Emily (17:13)
then you're getting absolutely nothing locally for this except the sort of short-term illusion of growth. And it's really scary what you say about how quick the turnaround can be. And if everything's being done behind closed doors and your first indication is the cranes coming in, that does make it harder to resist. And so think this comes back to, we've got to have transparency on this
Heidi (17:13)
then you're getting absolutely nothing locally for this, except the sort short-term illusion of growth. And it's really scary what you say about how quick the turnaround can be, and if everything's being done behind closed doors, and your first indication is the train's coming in, that does make it harder to resist. If I just come back to, we've got to have transparency on
Emily (17:33)
so that people have a say in what's coming into their communities.
Alex Hanna (17:37)
Yeah, and I think one of the go ahead, Joe.
Joel (17:37)
Well, and let me just add
this, and I'm sorry that I'm interrupting, but let me just add this. In the case of Fargo-Harwood area, which is separate from Allendale, but the same thing that happened. I'm not kidding. You had cranes on the ground building this structure within three, four weeks after the announcement, and the public didn't know about it. And so I apologize for interrupting, but it was that quick. And it's going on right now.
Alex Hanna (17:43)
No God.
think one of the problems, I mean, this is, I didn't know that it was that quick. That's absurd. One of the problems is that it's very difficult to get state and national leadership on this. ⁓ If you look at President Trump's AI action plan, the kind of guidance around data centers is build, baby, build effectively. It's just build as many of them as possible. And it is leveraging this argument around we need to beat China, we need to enhance science, all these different things, which are
not frankly not good arguments for the construction of these these different data centers for many different reasons. And so yeah, that's the leadership that we're seeing on the state level. Different states are approaching this in different ways. mean, even in places where you would anticipate ⁓ more a more ⁓ liberal leadership on this, you're not seeing this, you're not seeing this in California. ⁓ To my knowledge, the type of
⁓ The type of legislation that has been proposed has not really gotten anywhere. The only place we've seen a state level moratorium is Maine and the governor Maine vetoed it. then I know Cathy Hochul in New York has been weighing that, but that has to come with massive grassroots resistance leading that effort and really pushing her to do so.
Heidi (19:12)
vetoed it.
Yeah.
So let me throw another wrench into this. As somebody who has basically worked with a lot of utilities, as Joel has as well, you look at China, and if China is to be believed, they're developing chips that use a lot less energy, that are a lot less resource dependent. Okay, so where am I going with this? Think about a little local co-op.
which you have, it's usually public power in a lot of these rural areas, they're gonna build capacity to serve this, and then they're gonna have stranded capacity. And who's gonna pay that besides the payers who are left. And the other thing is you look at how many of these are on the books, and you look at is every one of those essential? And then you look at the risks that we have.
going forward in terms of stock market risk, equity market risk on basically on the back of ⁓ this construction, on the back of this hype, which you guys call a con. On the back of this, see record stock market returns only to think what happens when none of that materializes. So there's a...
there's not a lot of thinking about the what ifs. Now I'm not saying all of that's gonna happen, but it would sure be nice if public officials would actually have an honest discussion about what the risks are and how they're trying to mitigate those risks going forward.
Emily (21:09)
Yeah, and I think that there's a really nice contrast there or telling contrast in what you're talking about. There's all these promises that are fake. It's the hype. It's the con. I like sometimes to say instead of saying, AI, what if you said magic beans? I have some magic beans and they are going to revolutionize health care. I have some magic beans. And then you can see how silly it sounds and how ridiculous it is that we aren't asking more questions when people come in with these promises. And on the other hand, you have all of these really real impacts to roads, as you say, Joel, to the electrical grid to
Heidi (21:22)
Hahaha
Emily (21:39)
rate payers who have to handle that stranded capacity, and then to everyone's retirement accounts that are bound up in this bubble that's happening in the stock market. And I think it's really important to be thinking ahead to what we're going to do when this finally comes back down to earth because it's unsustainable and how we're going to make sure that we protect the people rather than the business interests that caused all of this.
Joel (21:48)
Mm.
Well, be careful because there's going to come a day when you two get to say, told you so. And nobody's going to want to hear it. But but let me give you an example of what's happening also here. Local leaders. I served in the state Senate and I still know some of these individuals that now are still in the legislature. They're getting run out of town. mean, locals are ticked off. These individuals are announcing that they're not running again.
in some cases, finding out that the the centers are going on land that they own, which shocking in a conservative state like this. But there's a price that people are paying for this politically. And I'm wondering if there's examples of where that's kind of changed things, where where you're seeing political leaders kind of come out and say, well, well, well, and I know you've given me some of those, but here's my point.
The knowledge, the knowledge that you're doing in your podcast, what we're trying to do here ⁓ using your knowledge. You know, I tried to get these guys on my show just and my criteria was on my radio show was they had to take cold calls. This wasn't going to be an hour interview. They were going to have to let people call in. They freaked. I mean, they were all OK until then. And of course, I was going to make them pay because I'm a capitalist and I'm not giving them the free time. ⁓ But
Alex Hanna (23:17)
Mmm.
Emily (23:18)
Mm.
Joel (23:30)
But they were so afraid, you two, of taking the cold calls. And I'm curious what access you've seen that the public has to them.
Alex Hanna (23:41)
⁓ Yeah, I mean, we've seen a lot of these different things when we've seen town halls and we've seen the public just completely go, what are you, you know, what are you telling us? What is, is this actually true? There's been representatives, I think, when there was a town hall or hearing in Memphis, which is really, really highly publicized case in terms of these X AI data center.
Heidi (24:01)
Yeah.
Alex Hanna (24:08)
in Box Town, which is this historically black community. they just, I mean, they ran out the XAI representative of the room. then, and had, mean, I don't know if there's been necessarily a legislative change or rather the actual change in office in Memphis, it would surprise me if there wasn't. It's surprising how much people are running on data centers now. I mean, in
rural areas in opposition of data centers. Really surprisingly, this person who led the effort in this township in Union County in Ohio where my mom lives, I think she lost her bid in the primary, her website was all around.
Heidi (24:39)
in opposition of data centers.
Joel (24:43)
Mm-hmm.
Alex Hanna (25:03)
no corporate welfare. And I'm like, wow, it's very surprising that a Republican would be saying this and be using that kind of language. And so I think we're going to see that change. mean, it's going to be very interesting what we see in the midterms and what we see how the house changes, especially when people are really concerned about things happening in their communities. And as y'all mentioned, it's happening on both sides of the aisle. It's going to be Democrats and Republicans who are kowtowing
Heidi (25:05)
Ha
Joel (25:07)
you
Alex Hanna (25:32)
to the industry for ⁓ different similar reasons.
Heidi (25:34)
or
right. I want to we don't have a lot of time left, but I want to switch from data centers to AI, you know, and I want to talk about hallucinations, you know, accuracies, you know, this idea that I don't know, anyone, not not a big fan always of South Park, but there's an episode where one of the dads is running a marijuana farm and he's going broke.
And so he falls in love with the, or he calls ⁓ AI, basically uses AI to give him some advice and ends up, ⁓ you know, it's an interesting because he develops a relationship with something that is not a person, which is how we started out this conversation about whether, you know, how enticing and how dangerous this technology is when we become too dependent on it.
But again, going back, you know, we got the data center. Yeah.
Joel (26:33)
Hold on, Heidi. Hold
on. You can't say that without telling us whether or not he sold more pot. OK.
Heidi (26:40)
No, he didn't. And in fact,
Alex Hanna (26:40)
Ha ha
ha!
Heidi (26:41)
in fact, the really interesting, I would recommend it. It's just so right. The really interesting thing is his wife gets really upset because he's just, he's spending all night, you know, talking to the AI bot. And so she gets jealous. So she starts training herself to talk like the AI bot. I mean, you know, so she could have a relationship with her husband again. I mean, you know, you think about how crazy that is.
Joel (26:50)
Okay.
Emily (27:00)
⁓
Joel (27:07)
You
Heidi (27:09)
But yet, we have this epidemic of loneliness. We have all of this kind of ⁓ attributing like a gender to this discussion. And I think that a lot of people are concerned, number one, about jobs, about what does this mean when you start replacing middle managers or lawyers, and how accurate can this actually be, and are we using judgment?
that are we looking at regulation, not just of the data centers, but of this technology and how this technology should be applied and used. Back to the initial discussion about people who bring this up in graduation speeches as it's the end all and be all are getting booed because these young people see this as a real threat to their livelihood. So what do you think this?
this is long way to ask this question. What do you think is the state of regulation of AI both nationally and locally and internationally? And have you seen any ⁓ attempts that you find particularly persuasive that you can recommend?
Emily (28:25)
Yes, so the first thing I want to say is that hallucination is not a great word, partially because it anthropomorphizes the system, right? To hallucinate is to perceive something that's not there and these systems can't perceive anything. But also because if what comes out of the system is right, that is also by chance. And so when we talk about them as having inaccuracies or hallucinations, we are sort of misconstruing what's actually going on in these systems.
Heidi (28:32)
Yeah.
Emily (28:52)
I think it's a very bad idea to use them even for information access because you are cut off from understanding the source of the information you're looking at. So I really think it's important to...
Heidi (29:03)
I
would say not always because if you use it enough it will reference the source and I have double checked the source. So depending upon what you're using.
Emily (29:09)
So if you
actually take the time to click through the source, then good. But at that point, you should respond to the source in the first place. And I think not everyone uses it as carefully as you do, Heidi. And if they see a source, they'll assume that means that that's where it came from. But that's not true. But that is also part of the general mush that's coming out of the systems.
Heidi (29:24)
Okay.
Yeah, well, I was going to say
Joel (29:28)
You
Heidi (29:28)
I don't trust it enough
to put my name on anything that I don't check the primary source.
Emily (29:33)
Good, good. So in the case
of lawyers, the very first time that one lawyer got in trouble for basically submitting something that he had extruded out of chat GPT as a legal brief, it was sort of understandable because these things are being sold as like all knowing everything machines, so why not? But of course it was inaccurate and he got in trouble. The second time it happened, I thought, don't lawyers gossip? Like how did that not show up as so embarrassing that nobody else would even dare? ⁓
Heidi (30:02)
Some people would say lawyers don't have the capacity to be embarrassed. And I could say I'm a lawyer.
Joel (30:06)
Hi, Heidi,
Emily (30:07)
Yeah, maybe that's it.
Alex Hanna (30:07)
Hahaha!
Joel (30:09)
you're a lawyer. So yeah.
Emily (30:11)
So, but I think that there is more regulation that's needed, but in many cases, just applying existing regulation is enough. So there was a win recently, I think it was in Pennsylvania, where it was held up that these things cannot be portrayed as psychotherapists ⁓ because they don't meet the definition that the regulations specify for what someone has to be to be able to provide psychotherapy, for example. And there was a recent law passed in Washington state sort of reaffirming
that they can't be called nurses because they don't meet the properties. So I think we can actually get pretty far with existing regulation if we're willing to actually apply it and not fall for the magic beans idea.
Alex Hanna (30:54)
Yeah, mean, regulation is kind of spotty. We're seeing, there's a lot of AI regulation that's being proposed. A lot of it has to do with non-consensual, like deep fake porn, the kind of effects it's having on minors, things that I think are pretty concrete harms. We're seeing kind of a spotty amount of regulation when it comes to, let's say, AI and employment. ⁓
there was in California something called the No Robo Bosses Act, which would have prohibited things like automated hiring and firing and setting of hours, which has passed the state senate and is on the governor's desk. And he vetoed it last session. Let's see what he does at this time on his way out. So we're seeing kind of different things. seeing we had a comprehensive piece of legislation passed in the EU, which
targeted certain types of ⁓ chatbot usage with regards to ⁓ risks, but that didn't really go far enough, especially in protecting people like refugees and other folks that don't have citizenship status in the EU. ⁓ But there are, as Emily says, lots of basic regulations that do exist, especially around data protection legislation. So in Europe, there's the general data protection regulation. There's a lot of things that can be done.
that can help push or prohibit ⁓ types of large language model usage in training, just in terms of if your data is used to train a model, you can ask it to be removed. ⁓ But that would actually, it takes a lot of work for them to do that. So enforcement's kind of hard. And then there's not comprehensive data privacy legislation in the US, but there are different types of legislation that we see in California.
in Illinois in certain places that have particular types of regulation. It'd be great to have federal legislation on this, we don't have a Congress that wants to do that right now. But there are types of schema that are the blueprint. And the thing that industry is going to say is, it's moving too fast. We can't regulate it. No, we have the tools. We have some of the templates. They're here now, and we need the political will to get them done.
Joel (33:14)
Well, I want to thank you two for coming on the hot dish again. Their podcast is a mystery. A I Hype Theater 3000. I think you owe it to yourself to check it out. You're spot on. I can tell you this living in the heart and soul of where some of this is happening. There are fights going on and it's going to get nothing but worse. But I thank you for coming on the hot dish.
Heidi (33:38)
Thanks guys. Good luck.
Alex Hanna (33:38)
Thanks so much for having us.
Emily (33:40)
Yeah, thank you and do keep
fighting, it's worth it.
Heidi (33:45)
So Joel, pretty interesting conversation and we kind of had to, they had a hard out so we had to let them go. But you know, it's interesting because ⁓ one of the OpenAI chat GPT, those folks say we're okay with regulation, but yet one of their founders has started a pack and spent hundreds of millions of dollars to prevent any kind of regulation.
So you've got to go back and take a look at where they're spending money. Don't listen to the hype of what they're saying about this, because are they really serious about having a conversation to meet the concerns of the people, never mind the data centers, but the concerns about AI? And they can talk a good game in terms of their public relations, but look at where they're spending money and who they're spending money to run against.
Joel (34:38)
Well, and the sad part to all of this height is there isn't and that's not like most of our guests and I'm not trying to diminish what they're doing, but there isn't much they said that I didn't know. ⁓ They nailed it when it comes to these data centers, as did you. ⁓ It's just that a lot of people listening to us on the hot dish right now don't know. And so I'm not trying to campaign on this platform as much as I'm trying to tell folks to get ahead of it.
Heidi (34:49)
Yeah.
Joel (35:07)
These non-disclosure agreements are disgusting. ⁓ The fact that this happened so quick in terms of construction is dividing communities. Heidi, you're right on when it comes to the rural electrics and what they're being set up to do. There's a fight going on amongst a rural water district. I know about just these types of things. It's all going to lead to fights down the road. So either government gets ahead of this.
and puts a process in place or a neighbor is going to be rolling on the ground with neighbor down the
Heidi (35:43)
Yeah, well, the one thing I would say is I am much more. I I I look into the future and and do not see. The kind of challenges that they see with using these products. I mean, I use them. I'm very careful about checking my sources, making sure if I'm going to cite a statutory reference, I check that statutory reference and so.
You know, these are really important tools and they're really one step beyond what you might say. What's the difference between that and a Google search? Well, a Google search isn't going to write your paper, right? It's not going to write your ABAD. It's not going to write your letter. A Google search is going to give you an outline that says these are some things that maybe you want to consider. But this is a really valuable tool. Where it goes wrong is when all of a sudden you have deep fakes.
where they can basically use our voices, our voices are out there, yours and mine, to basically steal our identity. The deep fakes in terms of CSAM, which is child sexually exploited material, where they can basically not use real children, but amalgamation of a lot of kids. And then, what are the child pornography laws related to that kind of... ⁓
⁓ image. And so there's just a lot of challenges right now with using these products, that the industry isn't willing to have that conversation and not willing to come to the forefront because by the way, they're running all the way to the bank. I mean, you know, Mark Zuckerberg, if we're going to believe that is is now $7 billion richer Joel today than he was yesterday. I mean, you know, so tell me how
how this is actually benefiting the economics of our country, middle class of our country, when they're taking away jobs and the billionaires are getting richer and richer and richer on the backs of building these data centers, wanting tax forgiveness for these data centers, because I guess they can't afford it. Think about that. The richest people in the world can't afford to pay taxes. And the whole while, they're basically funding candidates who don't want regulation.
Joel (37:57)
Well.
But, Heidi, aren't we really talking about two things? We're talking about AI, which is how it's being used and the fact that it can be used in such a positive way. I get it. I understand it. Since AI has increased ⁓ when I write anyway, my grammar is improved. I'm not trying to pretend that. ⁓ But let me just say this. We're talking about two things. We're talking about AI and we're talking about where you store it.
Heidi (38:18)
Yeah
Joel (38:31)
And those are two completely different things. And if you think that there's a danger of not needing it, you mentioned China in this argument. I think that's realistic. mean, here's another one, right? ⁓ Bitcoin. ⁓ When you look at what's happening with Bitcoin with these data storage units, ⁓ know, ask yourself whether not you think Bitcoin is a joke, but whether or not you think it's a currency of the future and
Heidi (38:47)
Yeah.
Joel (38:59)
Heidi, Bitcoin is a huge part of what's happening to us here in the Midwest.
Heidi (39:04)
Yeah, and it's it they're going places where they think power is cheap. It's not just Bitcoin, but crypto. so crypto mining, which is highly energy intensive, they're looking for cheap energy every place they can go. The more you put stress on the grid, the more you put stress on what's happening, ⁓ you know, for, ⁓ you know, utilization of electricity, the more you have to wonder who ultimately is going to pay the price. ⁓
Joel (39:12)
Mm-hmm.
Heidi (39:35)
as they kind of move out. I mean, I think I totally agree with you, Joel, but we're going to need more data centers. I can come, but we should not be giving it away. We should not be signing NDAs. We should not be not having a conversation about who's going to and how you're going to fund the resources. And the other thing that I would tell you is given
the number of these things that are on, it's like the gold rush, right? So you see everybody rushing to build a data center, trying to get financing, they're gonna overbuild. They're gonna overbuild because they don't wanna miss out. And so the question becomes, what does a good actor who is gonna build this infrastructure, what does that look like? And how should the public demand that that good actor, you know,
acquit themselves as they're building out this important part of the infrastructure. so I think going forward, the most important thing people can do is have a conversation locally and have a conversation with their state legislature. Because if you're expecting Congress to do anything, dream on. Congress is not going to get this done.
Joel (40:54)
Well, and that's the reason that I pointed out that it happens right away. Once I get going, it happens right away. And we're talking about a crane is bigger than any crane you've ever seen in North Dakota. ⁓ There's a million places to build these where they don't have the same effect as what they do, where they're building them now, right along the highway ⁓ in places where it's convenient for them.
versus convenient for the people around it. They got to deal with the noise. They got to deal with the roads. They got to deal with everything that goes along with it besides the utility needs. And so, yes, we may need them. But my big question as a former state senator is why are they doing them the way they're doing them when they don't have to? Yeah, and that's exactly what they're doing. Exactly.
Heidi (41:41)
Sneaking them in, sneaking them in.
So
Joel, I just have to say something because, you know, I look at you when we're doing this podcast, and I think it's so interesting because your head covers up one half of your poster and the other half has the Green Bay Packers on it.
Joel (41:59)
Well, let's look at what happens there. Let's look at Brett Favre getting sacked.
Heidi (42:06)
I'm just saying, I
just say it when your head is over the picture. I find it really, there you go, see, I can see.
Joel (42:12)
Well, how about now?
How about now? And can I just say this? I'm so sick of Republicans pretending that that flag isn't mine. I'm gonna keep it right there because it's mine too. So there. And the one separator between that flag and the Republicans flag that they're so proud of is that one never hit a cop. So there.
Heidi (42:18)
I'm sorry!
It's all of ours.
So, Joel,
you don't know this, but the White House called, they want you and I to sing at the 250th anniversary. They heard, they heard we've done funerals and weddings together.
Joel (42:40)
Nobody else is willing to do it.
Well, we have, and nobody else is willing to do it.
Heidi (42:48)
I'm just offering us, we
could go sing. Although you and I would not meet the apparently, there's a waist girth requirement to show up. You and I, although the president wouldn't meet it either, you and I may not pass the muster there.
Joel (42:54)
Actually, actually I think people.
Yeah,
I think that I think we got a better chance of two people going into the ultimate fighting ring, you know. mean, I mean, before you get off that subject, in fairness, how many people in your life are talking about or fired up about this nation's hundred or two hundred and fiftieth birthday? Nobody in my life is nobody.
Heidi (43:19)
Oh, it's so sad. It is so sad. mean, what?
Yeah
None. None. It's really, yeah.
Joel (43:37)
And we talked about it on my show today, which, by the way, it's called News and Views. But we we talked about it on my show. Yeah, actually, KFGO.com. But I mean, here's the truth. They've wrecked it. They've wrecked it. They've wrecked the birthday. And so they're the kid that came to the party and sat on the cake. So they're.
Heidi (43:43)
and it streams and you get podcasts as a result of it. Well.
Well, I mean, you can't look at what's happening on the White House grounds and what's happening with the reflecting pool. you know, it it it it what this should be a time for reflection on how important the United States is to the trajectory of democracy. You know, the beacon of the world and everybody's looking at us now internationally saying we don't recognize the people of.
the United States and people say, well, it's Donald Trump. I said, look, the first term it was Donald Trump. The second term it was the American people who put him there when they knew what he was gonna do. internationally.
Joel (44:37)
Well, let me let me throw
one in at your former colleagues who I know you've got friends on both sides of the aisle. ⁓ You know, nobody elected this guy King, and if Congress doesn't do its job and I mean the Republicans in Congress, then we've got one. So there I got that off my chest as well.
Heidi (44:49)
Yeah.
Well,
and think about this slush fund, ⁓ you know, which the judge just, you know, and that's the other thing that I will tell you that this is appalling to my Republican, former Republican, you know, me being a former senator, it's appalling to Republican senators, but they are too damn chicken. And they just love that the court has stopped it because they can say, let the courts deal with it. We don't want to ruffle any feathers. We don't want to take any risk. And so
You know, it's going to be interesting. This is an interesting week and you and I should probably have a conversation next week after we see what happens here with the reconciliation bill.
Joel (45:37)
Well,
you have you have always been the more proper and more decent of both of us. So let me just say this. They haven't been chicken. They've been chicken shit. So there I feel better having said that.
Heidi (45:50)
Yeah. Yeah, I watched an ⁓ explanation of what the judge said. Sheldon Whitehouse posted it on his Facebook page. It was really very, I sent it out to friends because they were saying, what did the court do? And it was really very well done. But he called it the cop beating slush fund, the cop beaters slush fund. And that's about the truth. Well, listen, we've got so much to talk about, but
Joel (46:11)
It's sad, but it's true.
Heidi (46:15)
You know, we thought it was important given what's happening right now in rural America. One of the biggest issues, big political issue is these data centers. What does that mean? And, you know, when we're talking about wealth inequality, we haven't really even scratched the surface of what the AI build out means for wealth inequality and, and what that means for the average middle-class workers. So,
Joel (46:37)
Yep, no
Heidi (46:41)
so much for joining us once again on the hot dish brought to you by One Country Project.
Joel (46:47)
You can learn more about us at onecountryproject.org. Be sure to follow us on Substack, YouTube, Facebook and Blue Sky.
Heidi (46:55)
And we're going to be back next week with more hot dish comfort food for rural America.