Ralph Nader Hour - Sunday, June 7, 2026

Follows the audio · click any line to jump

1. And seven FM Los Angeles and online at KPFK dot org.
2. Yo, this is your brother's speech from the crew arrested development.
3. You rock it with KPFK 90.7 Los Angeles.
4. This is Chris Hedges and you're listening to the Ralph Nader Radio Hour.
5. Welcome to the Ralph Nader Radio Hour
6. My name is Steve Scrovan along with my co-host David Feldman
7. Hello David
8. Hello Steve
9. And of course as always our producer Hannah Feldman
10. Hello Hannah
11. Hello Steve
12. And it's not a Ralph Nader Radio Hour without the man of the hour
13. Ralph Nader, hello Ralph
14. Hello everybody
15. All right, so the gang's all here for today's program, where we are going to do a deep dive into AI.
16. According to a recent Gallup poll, 7 in 10 Americans oppose constructing data centers for artificial intelligence in their local area.
17. This comes as no surprise to groups like Public Citizen, which has been studying the impact of these massive data centers on everything from consumers' electric bills to the global climate crisis.
18. Tyson Slocum is director of Public Citizens Energy Program,
19. and he's going to lay out the threats these data centers pose
20. and how communities across the country are already resisting.
21. Then in the second half of the program, we welcome New York Times reporter David Wallace-Wells,
22. who in a recent New York Times Magazine piece titled,
23. AI Populism is Here and No One is Ready.
24. He argues that tech oligarchs like Sam Altman and Dario Amadei,
25. anxious about their technology taking over the world,
26. didn't count on a backlash from real live human beings.
27. Will the people armed with pitchforks and Molotov cocktails
28. rise up against this small band of billionaire tech bros
29. before the machines do?
30. As always, somewhere in the middle,
31. we'll check in with our tireless corporate crime reporter,
32. Russell Mokhyber.
33. But first, resistance to data centers is growing.
34. David?
35. Tyson Slocum is director of Public Citizens Energy Program,
36. covering the regulation of petroleum,
37. natural gas, and power markets. He serves on the U.S. Commodity Futures Trading Commission's
38. Energy and Environmental Markets Advisory Committee and frequently intervenes before
39. the Federal Energy Regulatory Commission, representing the interests of household
40. consumers. Welcome to the Ralph Nader Radio Hour, Tyson Slocum. It is my pleasure to be here. Thank
41. you. Welcome indeed, Tyson. Listeners, this is going to be on data centers and the huge
42. opposition all over the country. But I want to start from the very beginning with Tyson. So we
43. clarify all kinds of questions that the press does not focus on because they're focusing on
44. where the data center is being built and who's opposed to it and who's pushing it.
45. So let's start with this. Name the major companies who are building these data centers.
46. The largest single owner and operator of data centers is a company called QTS, which is a portfolio company of the private equity giant Blackstone.
47. So what a lot of people don't realize is that most data centers are actually not built by the
48. big tech conglomerates like Meta or Alphabet or Microsoft, but by real estate development firms
49. that are primarily backed by private equity But clearly the big four from Silicon Valley are the initiators They the ones that got the money You want to name them
50. Google, Meta, Microsoft, Oracle, NVIDIA.
51. Those are the big ones.
52. And these data centers now are provoking a tremendous opposition.
53. I've never predicted it.
54. People are worried about higher electricity prices because they devour a huge amount of energy.
55. They're worried about loss of water sources because they devour a huge amount of water.
56. These data centers, they don't provide many jobs.
57. They have noise.
58. They take up valuable land and they manipulate politicians.
59. They give campaign money.
60. They cut sweetheart deals.
61. Some of them are getting exemptions from sales tax, exemptions from property tax.
62. Some of them, not all of them.
63. It depends on how vigorous the local and state governments are.
64. Give us an idea how much a data center is costing, where they're getting these enormous
65. hundreds of billions of dollars a year that they're starting to spend, these Silicon Valley
66. companies.
67. Give us the terrain, Tyson's Flocom.
68. Yeah, so data centers are becoming sort of the physical presence of how a big tech company is able to have its market share of AI recognized.
69. So a lot of AI applications at this point are theoretical.
70. They are promised in the future that they are going to be replacing entire industries within our economy.
71. And so, as you know, Wall Street and the equities market values not just what a company is doing today, but more importantly, how they're positioned in the future.
72. And what Wall Street and investors and what big tech companies are selling is this vision that AI is going to replace everything else.
73. And that AI is going to be monetized, that you are going to have to subscribe and pay for their intelligence that is going to be replacing your kids' jobs and everything else.
74. And so because so much of AI application from a market perspective is speculative, the primary way that investors are measuring how a big tech company is performing in dominating market share is what is your data center capacity to produce generative AI and store and manage this data.
75. And so data centers have become the sort of steel in the ground measuring stick of how to determine whether or not a big tech company is dominating in artificial intelligence market share.
76. And so, OK, let's break it down.
77. How big are these data centers?
78. How many are there already operating?
79. How many are under construction and how much money goes into building these data centers?
80. Give us some data on the ground, so to speak.
81. Yeah, so there are hundreds of data centers that are located all over the country.
82. And we've been building and operating data centers for a very long time.
83. When you take a picture on your Android phone and you store that picture in the cloud, there's a data center.
84. There's a data center needed to stream Netflix and other entertainment services like that.
85. What has been different is that the computing power necessary to support generative AI and large language models is so many orders of magnitude larger than all of the existing data centers That why it only been in the last two or so years that it become a super controversial
86. thing because now there are hundreds of proposed data centers under construction or being proposed
87. in every U.S. state. And whether or not these are going to be built is really the key fight
88. that we're seeing at the state and local level right now. How many square feet is an average
89. data center? It really depends. You've got smaller scale data centers that can just fit inside a
90. normal size warehouse. And then you've got so-called hyperscale data centers, which there
91. are none in current operation, but several proposed and under construction. And those are
92. typically defined as having computing power that requires at least a thousand megawatts of
93. electricity in order to supply the energy for the microprocessors to produce that artificial
94. intelligence. So it really runs the gamut. We've got a lot of small scale ones that are tucked
95. away into a commercial or industrial neighborhood and you've got much larger ones. So it really
96. depends. But again, the big tech companies in a race to control perceived market share keep
97. announcing larger and larger data centers that really is just about public and market jockeying
98. for perceived market share. And then in addition, Ralph, there are a lot of data center proposals
99. that do not have a client secured yet.
100. So they don't have a big tech client.
101. And this is because as Wall Street creates a massive speculative financial bubble for
102. big tech valuations, you're now seeing a corresponding real estate speculative bubble
103. as developers seek to cash in on the AI tech bubble by buying up land.
104. A lot of times it's farmland and they do it under an LLC, secure initial permits, and then go out and try to find a client to lease the facility to.
105. Are the contracts between these data center companies and state and local government public?
106. Have you seen them?
107. Have you got most of the time?
108. No, most of the time.
109. So what we have seen are a proliferation of what's known as nondisclosure agreements.
110. That's when a project developer comes to the local community and says to the mayor or to the zoning board or to the County Economic Development Commission and says, listen, we've secured rights to land in your area.
111. We are considering building a data center. We are also considering other sites for this data center. Because it is such a top secret project, we're going to require you to sign a nondisclosure agreement that prevents you from sharing any information or detail about this project.
112. And because these nondisclosure agreements are so pervasive throughout the country, reporters, local residents cannot get necessary details about the project.
113. And that's really been one of the reasons for the big backlash.
114. Are there any Freedom of Information Act lawsuits filed against the state and local governments for keeping this information secret underway right now?
115. Yes there have been a number of those And also in response to documented abuses we have seen municipalities counties and even states move to change the laws So in the Phoenix area for example where we had documented abuses
116. of nondisclosure agreements between Amazon and the project developer and the local community,
117. there is now a law in the books that immediately releases all signers to the nondisclosure
118. agreement at least 90 days prior to any public hearing about the proposed project. And so what
119. that does is that enables the local community, local journalists and others to be able to have
120. unfettered access to those underlying documents prior to any public hearing where the merits of
121. the project would be debated. So the problem is right now, Ralph, is we are looking at massive
122. stranded assets. Because when you build a data center, you're not just building a standalone
123. data center. You are building billions and billions of dollars of utility steel in the
124. ground infrastructure that has a life of 40 to 50 years. And with the Trump administration,
125. most of that infrastructure is fossil fuel or natural gas. So that's the big problem.
126. So I definitely see that we are in a speculative bubble.
127. That bubble will burst.
128. And folks within the AI industry, like Sam Altman, have been very clear where they have
129. publicly said, when the bubble breaks, we expect to get a financial bailout because
130. our AI applications are so important to the national interest.
131. Tyson, let's parse that.
132. Traditional market structures always experience overcapacity, otherwise known as boom, bust,
133. you know, prosperous times, recessions, depressions, et cetera.
134. What you're describing seems to defy this limitation of capacity.
135. Do these data centers adhere to the marketplace dictate of a time when there is going to be
136. over capacity, some data centers and companies have to close and others in a better position,
137. better managed, perhaps thrive? Or are they creating a new kind of market fundamental
138. determinism that is infinite? Well, I think we are already at the sort of physical peak of
139. data center capacity because there is just so much copper wire, transformers, electrical components,
140. and power generation equipment that can accommodate the data center boom, right?
141. So we're already bursting at the seams and running into shortages.
142. And data centers aren't just causing electric rates to go up because of their power demand
143. issues, but the fact that big tech is gobbling up all of the needed infrastructure and components
144. that the rest of the power grid requires.
145. And so prices are going up for every type of component.
146. And, you know, Trump's punitive tariffs aren't helping with those global supply chain bottlenecks.
147. You're talking about supply here.
148. I'm talking about demand.
149. Is there going to be unlimited demand to provoke unlimited supply demands of materials to create more and more data centers?
150. So what I'm saying is big tech envisions an infinite, unlimited demand scenario, but there is a physical limit to being able to build the data centers that big tech claims they need.
151. But going back to this first point of is this all encompassing dominant replacement vision of artificial intelligence going to happen?
152. The voices on the ground, Republican, Democrat, it doesn't matter, are emphatically saying no.
153. And the backlash to data centers isn't just about, oh, I'm concerned about my power rates going up or I'm concerned about the noise or the water usage.
154. it's also a civil rights and human rights issue where people are saying,
155. I don't like this vision that big tech is laying out for us that is going to be produced in this
156. building down the street from our community. What you're describing, Tyson, is basically
157. a formidable sub-economy where the sellers create their own demand. The demand comes from the
158. sellers. They're basically dealing with themselves. But Wall Street, Ralph, buys into that vision.
159. That's why there's five tech companies that represent a third of the entire market capitalization of the entire S&P 500.
160. So five companies have 33% of the value of the 500 biggest companies.
161. That's because Wall Street is buying into this vision that AI will replace everything else.
162. all other companies in the portfolio and AI and the big tech backers have this draconian vision
163. of AI replacing labor, of becoming the dominant feature in our economy,
164. defining interpersonal relationships. And a lot of people out there, a lot of listeners out there,
165. Ralph, have had interactions with AI, have read the warnings from the AI developers themselves.
166. Here's my concern, Tyson, that we have now allowed this technology to reach a point of no return.
167. That basically, its political power, its technological dominance, its rapid development almost by the month with new revelations of scary AIs has become what the dean of Harvard Law School a century ago, Roscoe Pound, once called beyond the effects of legal action.
168. that you're not going to get regulation, you're not going to get statutes. There'll be a few
169. lawsuits here and there on behalf of severely damaged teenagers, as is already occurring
170. under tort law. But basically, it's out of control. And the Congress is sitting there
171. having hearings, pounding the table with questions, both, as you indicated, from Republicans and
172. Democrats. But the bottom line is nothing has happened. It didn't happen in 2005. It didn't
173. happened in 2010. It didn't happen in 2015 in Congress, didn't happen in 2020. And there's no
174. indication that any bill is heading out of committee and going to the floor. What's your
175. reaction to all this? Is it beyond the effects of legal action? It cannot be controlled.
176. Absolutely not. So at the federal level and at Congress, you know, Congress has been completely
177. inept and the Trump administration seeks to accommodate whatever big tech wants.
178. So did Biden administration.
179. Right. The Biden administration was a problem as well.
180. Not as overt.
181. But at the state level, Ralph, we are seeing huge pushback. And this is why the Trump
182. administration last year publicly pushed and tried to enact into law a 10-year moratorium
183. on any state law regarding artificial intelligence because the Trump administration aligned with big tech is trying to stop states from regulating artificial intelligence content to protect kids
184. from predatory practices, and to adequately regulate the energy and resource uses for data
185. centers. So Trump and big tech are trying to spin a tale that the AI revolution is inevitable,
186. and there's nothing to do to fight it. But what we're seeing at the ground level, Ralph,
187. is bipartisan outrage at this vision that none of us consented to and that none of us want.
188. And my fear is that the Democratic Party, which historically has been aligned with big tech and
189. relied upon big tech to finance their campaigns, some corporate Dems are panicking and seeking to
190. try and accommodate big tech to try and pry them from Trump and the Republicans. And so I think
191. the other civil war that's going on here is within the Democratic Party. Is the Democratic Party
192. going to listen to the people, to the base, who say, we do not want this draconian, all-encompassing,
193. controlling vision of artificial intelligence and a data center in every neighborhood? Or are they
194. going to go with what big tech and their trillions of dollars want? And I think what we're seeing
195. right now, Ralph, is huge, sustained and well-organized opposition in deep red, deep blue
196. districts. It really doesn't matter. And the thing that is unifying all Americans in 2026
197. is opposition to AI and data centers. Let me interrupt here. Let me counter the
198. obsolescence of your declared hopes, which our listeners have just heard. This is nothing that
199. can be stopped at a state or local level, except very, very minimally and very limited geographically.
200. This is a global field of action by the Silicon Valley fanatics. They're operating globally
201. because AI operates globally. It's very hard to sue AI under tort. Who do you sue? Where do you
202. sue? It's everywhere. It doesn't recognize national borders. So you can have a spurt of
203. defiance and regulation in this state and that state. And it's good to keep the fires of revolt
204. and resistance alive, but it's only nibbling at the edges. We have to start nationally here.
205. We have to federalize the charters of these giant corporations and rewrite the compact with these companies and society from the early 1800s when the modern corporation was chartered by legislative action, the textile companies.
206. all obsolete. We have to get these giant companies out of their chartering jurisdiction of Delaware,
207. Nevada, and now Texas. And the basic question is, they are developing essentially governmental
208. powers, governmental powers, not market powers or corporate powers. They've reached the level now
209. where they are our government, the corporate governments. And we have to escalate our
210. urgencies to that level. It's more than just the hour is late. The hour is over. So we have to
211. go back and respond with a completely unprecedented level of public interest standards, etc.,
212. including whether this technology should be allowed at all. Remember, the supersonic
213. transport planes were considered inevitable and they flew from Paris and London to Dulles Airport but eventually they were blocked
214. And we don't have supersonic passenger planes.
215. It's not often that if technology is blocked, we have to start thinking of that, that it is basically unable to adhere to the health, safety, and economic rights of a civilized society.
216. It's transcending them. And even their promoters are getting scared to it. They're starting to
217. talk about a universal basic income for the proletariat that is now completely unemployed.
218. You think your job is going to be taken in 10 years by AI?
219. I certainly hope not. But Ralph, I'm not asking you whether you hope. Do you predict?
220. I don't think my job, but a lot of jobs, if big tech gets its way. I agree with everything you
221. said, Ralph, and that's why we need to break up big tech. They are engaged in too many activities
222. and dominate too many of those activities. It's an antitrust issue. Their political and economic
223. control is way too concentrated. We need to get Lena Kahn back into running the Federal Trade
224. Commission and going after big tech. Tyson, you're obsolete. This is nothing that can be
225. dealt with by antitrust. You're just breaking up big monsters into smaller monsters who then
226. become big monsters. Forget about that. You better have a roundtable discussion with people
227. who can basically predict the rest of the century here as best as they can. Forget about that
228. antitrust. Now, Bernie Sanders wants a moratorium on these data centers, and you disagree. You're
229. too modest. You're going to be overwhelmed by the reality. Why are you opposed to Bernie Sanders?
230. I'm not opposed to Bernie Sanders. What I said was I am opposed to a federal moratorium on what
231. is traditionally a state and local issue. If a local community wants to support a data center
232. in partnership with the local hospital that is going to be 100% run on solar power with
233. on-site battery backup storage to come up with, use AI to help cure leukemia, under Bernie Sanders'
234. bill, that would be prohibited. It prohibits all data centers for any use until there is
235. comprehensive federal reforms. My only point was, I'm seeing a lot of reforms being pushed at the
236. state and local level, and that a blanket moratorium on all data centers, I think,
237. is counterproductive because... I think Bernie's referring to giant data centers, not adjacent data
238. centers to hospitals. Tell our listeners how they can access you and Public Citizen on this subject.
239. Of course, Ralph. We're at citizen.org. The group is Public Citizen, and I'm Tyson Slocum,
240. the director of the energy program. Tyson, a curious last question I'd like to put to you.
241. Have you visited a data center, a big data center like the ones in nearby Virginia? And are these
242. data centers allowing civic tourists? I mean, even nuclear power plants have been known to allow
243. some visitors to look over the place. I am not aware of any data center in the United States
244. that has open houses where the local community
245. or others can tour.
246. Instead, most data centers look like
247. a medium security prison.
248. It's a large windowless building surrounded
249. by a security fence perimeter.
250. I spoke at a data center conference last year,
251. an industry conference and a gentleman introduced me to a French company that develops data centers in France where once a month the data center is open to the public They have a community meeting room where members of the community are informed what types of
252. applications and platforms are being developed at that facility.
253. They have apprenticeship programs, not just on the electrical side of things, but on the
254. tech side of things.
255. In the United States, that's not been the approach by big tech or by data center developers.
256. Well, we're out of time.
257. We've been speaking with Tyson Slocum, who heads the energy section of Public Citizen.
258. And this is obviously a subject to be continued, Tyson.
259. Thanks for coming on and exchanging with us for our listeners.
260. I'm sure we're going to get a good reaction to this program.
261. And as we close, give the website once more, and we'll be moving to our next guest.
262. Thank you, Ralph. It's citizen.org. And thanks so much for having me.
263. You're very welcome.
264. We've been speaking with Tyson Slocum. We will link to his work at publiccitizen
265. at ralphnaderradiohour.com. Now let's take a short break. When we come back,
266. we continue our exploration of the backlash against AI with New York Times reporter David
267. Wallace-Wells. But first, let's check in with our corporate crime reporter, Russell Mokhyber.
268. From the National Press Building in Washington, D.C., this is your Corporate Crime Report
269. of Morning Minute for Friday, June 5, 2026.
270. I'm Russell Mokhyber.
271. The White House pushed health officials to re-examine the Food and Drug Administration's
272. repeated rejections of an advanced melanoma medication from RepliMune.
273. That's according to a report in the Wall Street Journal.
274. The FDA had declined to approve RepliMune's treatment in July last year and then again
275. in April saying the data from the trial in which all patients received the experimental drug with
276. no comparison group fell short of proving to the agency that it worked. The FDA also said the mix
277. of patients enrolled made the results hard to interpret. After its application was rejected
278. for the second time, Rep. Lemieux and representatives met with the White House in early May
279. arguing the rejections didn't match up with the Trump administration's desire to help
280. terminally ill patients. For the Corporate Crime Reporter, I'm Russell Molcott.
281. Thank you, Russell. Welcome back to the Ralph Nader Radio Hour. I'm Steve Skrovan,
282. along with David Feldman, Hannah, and Ralph. Tech oligarchs predicting the rise of machines
283. may not have counted on the rise of the people. David?
284. David Wallace-Wells is a columnist and staff writer at the New York Times,
285. where he writes a weekly newsletter on climate change, technology, and the future of the planet.
286. is the author of the book, The Uninhabitable Earth, Life After Warning,
287. a travelogue of the near future and meditation on how it will look to those living through it.
288. Welcome to the Ralph Nader Radio Hour, David Wallace-Wells.
289. Thanks so much for having me. Great to be here.
290. David, you know you're on the frontiers of warning what's coming in the future.
291. You don't sugarcoat the techno twits promises of great abundance and freedom and liberty and prosperity.
292. from generative AI. So could you tell us what you meant when you said the AI backlash will be ugly
293. and no one is ready, not even tech CEOs? Well, just over the last six months, there's been a
294. huge surge in anti-AI and in particular, anti-data center organizing and activism in the US. And you
295. can see that on the ground where you see huge crowds coming to town halls to protest new data
296. data centers that are being proposed.
297. You see some towns that have approved those data centers,
298. literally having their entire city council voted out of office as a result.
299. And you see it in these surveys where within the span of just a few months,
300. huge sentiment flips among the American public from being basically agnostic about AI with some misgivings and some optimism
301. to pretty striking majority opposition to the technology and the infrastructure build out that it requires.
302. Now, there's a lot of stuff that goes into that opposition.
303. Some of it is environmental. Some of it is they just don't want this stuff being built in their backyards.
304. Some of that is about electricity prices that are going to be driven up to some degree by data center usage.
305. Some of it is about the ultimate impacts of the technology and what it will mean for employment and the future of the American economy.
306. But I think underneath all of that, and that really at the core, is this understanding that anyone with their eyes open can see pretty clearly that this is a technological revolution that has been designed and is being built by an extremely small number of people with very particular idiosyncratic, in certain ways, I think, somewhat sociopathic worldviews.
307. views. And in that sense, it doesn't play like a new industrial revolution to the American public.
308. It looks more like an intensification of the patterns that we've observed over the last
309. couple of decades of intensifying income inequality and growing signs of genuine plutocracy. And that,
310. I think, is the core reason why there's been so much public unrest and distrust and backlash.
311. It's not just that we're uncomfortable about the technology and what it means for our own futures.
312. It's that we see that this is a technology that is being controlled by a very small number of extremely rich, extremely powerful people who are essentially operating outside the realm of democratic control.
313. They're starting to get worried themselves.
314. You have your article opens with a quote, I prep for survival.
315. Open AI, Sam Altman confessed back in 2016.
316. I have guns, gold, potassium, iodide, antibiotics, batteries, water, gas masks from the Israeli Defense Force and a big patch of land and Big Sur.
317. I can fly to end quote.
318. So they're getting worried about a backlash, although they don't like to talk about it.
319. But also, you know, Elon Musk was written up, I think, in the New York Times recently saying that maybe there should be a universal income plan for all Americans who are going to be unemployed in the tens of millions by generative AI.
320. What's the background here?
321. Are they trying to look ahead here?
322. Well, I think, you know, the quote that you read from Sam Altman in 2016, it's from a slightly different phase of his life and his sort of anxiety trajectory.
323. That was at a time when many people who are working on AI were really worried about what they call the existential risks involved, which is to say, these are not scenarios that I personally put all that much stock into, but they have animated the worldview of many of the people who have built these systems.
324. They believe that once AIs were able to surpass humans in cognitive ability, pretty soon we'd find ourselves in a situation where that machine intelligence was running the world, possibly not to the advantage of humans.
325. So they worried a lot about what would happen if there was a kind of a Terminator scenario or a Matrix kind of scenario.
326. They built that technology anyway, and they didn't even really stop to think along the
327. way that if we were building this world-changing technology, which had dramatic implications,
328. some of them positive in the realm of say cancer research or whatever but plenty of them that are quite concerning or at least nerve that if we built this technology talked openly about how dramatic and consequential its rollout would be sort of casually remarked about what life would be like after there was no
329. need for human employment anymore.
330. They didn't even consider that that would produce any anxiety in the general public
331. and didn't really do much to protect themselves or their companies against that public backlash.
332. They kind of took the basically the sort of the Uber approach where they decided they were just going to ignore public resistance, ignore regulation, barrel ahead and trust that empowered with so much money that they could overcome any obstacles that the public threw up in their path.
333. Now, over the last year or so, I think they've started to change their perspective on that. They've seen a lot of this polling. They've seen the resistance to the data center build out. And they've started to talk a different tune.
334. many of these CEOs who are kind of constantly talking on podcasts and on panel discussions
335. and conferences, stop talking casually about what it will mean when there's no white-collar
336. jobs anymore and started talking much more frequently about what they might do to ease
337. the transition to that post-employment future.
338. They talk about universal basic income.
339. They talk about the importance of redistributing the gains from AI.
340. They talk about a new deal for AI.
341. And lately, the last couple of weeks, they've been talking a bit more also about how much philanthropic money will be coming out of the AI boom, which is interesting in its own right, since so much of that money is being directed towards AI safety, not to the traditional goals of developmental aid like philanthropies in the past.
342. But whether that any of that talk is sufficient to calm public anxieties and whether they will actually feel obligated or compelled to try to make any of that happen, I think, is very much an open question, which is maybe that's even putting it too generously.
343. I suspect that these figures, these leaders, these companies are not especially interested in securing public consent for a technological revolution.
344. They simply want to be given the space today to operate as they would like to.
345. And I think it's notable, you know, we're talking this week, the end of May, it's the same week that the Pope released his encyclical on AI, and which I think is a kind of significant document in our ongoing effort to understand the future of this technology.
346. But what I would emphasize now in this context is simply that that encyclical seemed to almost enter into a vacuum, which is to say it is not the case that there has been large scale coordinated, reflective opposition to this technology in our political sphere, in our cultural sphere.
347. there is this grassroots rebellion that we're seeing, but there's simply nowhere obvious for
348. that to go because so many of the leading institutions and organizing arenas of our
349. public life have abandoned the responsibility to really oversee, regulate, and shape the forces of
350. kind of techno-capitalism that have been unleashed here. And I think that's really unfortunate.
351. That's one reason why the Pope stands out as such a singular figure here. It's because there are so few other people who are singing the same tune and speaking to all of the Americans and all the people around the world who may be made at least uncomfortable by this technology.
352. It's interesting. Years ago, there was more outcry by the scientists and technologies.
353. You probably remember the 2014 letter to the world by leading scientists and even Elon
354. Musk as a technologist signed it warning that the robots can go out of control and take over the world And it could be a global disaster for humanity Earlier than that around 2002 Bill Joy a Silicon Valley type
355. wrote an article for Wired Magazine called The Future Doesn't Need Us. Again, talking about a
356. robotic-dominated future. There have been congressional hearings with a lot of table
357. pounding, but really no prospect in the foreseeable future of any comprehensive regulatory
358. legislation. But what do you see? What do you see at the state and federal legislative level?
359. Yeah, I think there were a few hearings back when Chachipetit first came out in late 2022,
360. early 2023. And there's been tussling over these questions in Congress and at the state level ever
361. since, but nothing significant or comprehensive. And I think the basic story there is that the
362. group in Silicon Valley who may be described as accelerationists, who think that the best thing
363. that we can do with AI is to unleash it, has sort of carried the day, especially under the Trump
364. administration. A lot of it has to do with money. These are now companies that are spending an
365. enormous amount of money on lobbying actually more than crypto, more than the fossil fuel business.
366. They've been the biggest donors in this upcoming political cycle, even though the threat of
367. regulation has been, to this point, relatively minor.
368. But there's also a kind of an ideological, intellectual component to the story, too.
369. And here, I would say I was really struck recently watching a conversation with Dario
370. Amodai, who's the head of Anthropic, which is generally understood to be, of the competing
371. leaders in the AI field, the most liberal, the most humane, the most humanistic in its
372. approach.
373. Each of these companies was founded because people who worked for a previous company got
374. alarmed at what the people who were running it were doing and ran off to start their own.
375. Anthropic is the sort of most admirable of those companies.
376. And in general, I think that Dario Amorai, he has many faults.
377. Of all of these leaders would be the person who I would most trust to be thinking thoughtfully
378. and conscientiously about these questions.
379. And I saw him in a conversation where he was saying that he didn't believe that ideology could survive the arrival of this technology because he thought it was too powerful.
380. Its reasoning powers would be too strong so that any debates, for instance, about the proper allocation of resources, which used to be contested in the realm of politics, would be resolved at the level of abstraction by machine intelligence.
381. and that we very quickly would learn to simply defer to those judgments.
382. Now, I think that's an extremely dangerous intuition, especially coming, as I said, from the member of this crew,
383. who I think is the most humanistic and thoughtful.
384. I was going to ask you about Anthropic, because in your article, you talk about this company, Anthropic,
385. which is probably going to go public soon, refusing to release, in your words, a new model, Claude Mythos,
386. which the company said could find and exploit security vulnerabilities in every tested piece of software, including those used in critical pieces of global IT infrastructure, end quote.
387. And that led to a serious conflict with the Pentagon and with the Secretary of Defense, Pete Hegsess, self-styled warrior of God, to berate and threaten Anthropix contracts with the Pentagon.
388. Can you explain clearly to our listeners what that was all about?
389. These are actually two different stories which happened in sequence, and I think
390. it interesting that they did happen that way So CLAWD is Anthropix leading LLM model It basically understood to be at the frontier ahead of every other model that been produced by any of these companies So if you
391. wanting to use absolute frontier cutting edge tech, Claude is your bet. And as a result,
392. Anthropic had entered into an agreement with the Pentagon to license its models to be used
393. for military purposes going back, I think they started this agreement about a year or so ago.
394. And that was notable in its own right because Dario is in many ways a kind of pacifistic figure.
395. Now he has, like many of these people in Silicon Valley, a strong sense of rivalry and competition
396. with China. So you could also describe him in a certain way as a China hawk. But again,
397. when compared to Elon's AI, for instance, Anthropic and Dario's organization are much
398. more humane, humanitarian. So the fact that they were doing business with the Pentagon was quite
399. notable. Now, at some point, the Pentagon came back to them and said that some of the terms
400. that were embedded in that agreement, the Pentagon wanted to revise. And these were terms concerning
401. limitations, what the models could be used for. Anthropic had previously insisted that their
402. models could not be used for purely autonomous use of weapons, which is to say, attacks on other
403. nations or their armies without any human involvement in the decision making.
404. And they wanted to limit its use for surveillance against American citizens.
405. And the Pentagon said, we need to revise those conditions.
406. It wasn't clear whether they had particular uses that they wanted to put Claude to that
407. would have violated either of those terms, or they were just looking a little bit further
408. down the road and thinking in two or three years, we might want to be able to use these
409. tools for these purposes.
410. Let's clear this up now.
411. But Anthropic balked and said, no, we're not going to do that. And as a result, not only did they have their contracts canceled, but they were actually declared to be a threat to national security, which has all of these downstream consequences, which are probably too easy to get into now.
412. But for the moment, suffice to say, it would mean that anyone else doing business with
413. Anthropic and Claude could also be prosecuted for actions against the U.S. government.
414. Now, that action was sort of withdrawn, but Claude remained an actor in sort of bad standing
415. in the national security space and in the Trump administration.
416. And then they had this further breakthrough where they came up with this new model that
417. wasn't trained to detect vulnerabilities in IT infrastructure, but was incredibly good
418. at doing so such that, as you say, basically everything they tested against, it found all
419. of these vulnerabilities that hackers could theoretically exploit, that foreign countries
420. could theoretically exploit, which had never been found before. And in a sense of obligation and
421. sort of responsibility, the company basically said, we're not going to release this model
422. because doing so would empower all these bad actors around the world who would immediately
423. be able to take down power plants and crash banking sites. And instead, we're going to share
424. it with a small number of critically important companies so that they can use the tool to find
425. those vulnerabilities in their own infrastructure and fix them. And maybe at a later date, we'll
426. release it to the public. But only once we've satisfied ourselves that all of this globally
427. critical infrastructure was secured. So that's unfolding. I think as a result, Anthropic is now
428. somewhat more in the good graces of the Pentagon. The whole story does illustrate this kind of
429. bigger question, which is both practical and philosophical, which is about how do we integrate
430. machine intelligence and artificial intelligence, LLMs, into military decision-making?
431. I think one thought experiment puts us into a place where we imagine machines making decisions
432. completely autonomous.
433. anonymously about who to target and where to go. But we're pretty far from that. And for the moment,
434. there are still humans involved in what's called the chain of command. But nevertheless, we are
435. seeing already in wars that we're fighting today, a strange kind of phenomenon where, for instance,
436. the US military on the first days of the Iran war targets this girl's school in Iran, kills 150 or
437. 200 girls and their parents who are coming to the school after the first strike to rescue their
438. children, only to be killed by a second double-tap strike, that rescue mission. And we know that that
439. target was chosen with the use and help of AI tools, including Claude. Now, in normal times,
440. we might say, okay, who authorized that strike? Who was the person who made the ultimate decision
441. that that school was a worthy target? We would have articles about it. We would have court
442. marshals about it. That person would be brought to some kind of military justice in time.
443. Instead, we've had this weird story that this target was somehow chosen by artificial intelligence and therefore is a kind of a cloud, a fog of war mistake being made by this new technology.
444. And we haven't had any real reckoning with who was in charge of applying that technology, who approved the choice of target, and who was ultimately responsible.
445. And I think this illustrates a big challenge with AI generally, which is that in this in-between period, when humans are still using these tools and are commanding these tools, are responsible for these tools, but are deferring to them to some degree, we are treating the tool itself as the decision maker and absolving the humans involved of their ultimate responsibility.
446. On that point in your article, you offer chilling prophecies, you call it. There was a conference at Yale and Dean Ball, the policy wonk who was an architect of the Trump administration's original AI policy, said, describing AI as, quote, this giant acid vat, end quote.
447. He said, it will not be AI and government, he predicted. It's going to be AI as government. Those are his words. It's going to be AI as government.
448. And earlier in your column, you say that this is the most personalized sales pitch ever foisted on the passive American consumer, a vision of a near total takeover of the country's economic, social, and cognitive lives by tools engineered by just five companies run by five particular people, several of whom are widely described as sociopaths, end quote.
449. Those are your words.
450. The reader of this article is entitled to ask, who are you describing as sociopaths?
451. Well, I guess I would say Sam Altman and Elon Musk have both been described that way by many people who've worked with them over the years.
452. So that's the head of OpenAI and the head of XAI.
453. They were both involved just recently in a lawsuit against one another in which a lot of this came out.
454. But much of it also appears in great reporting that a writer named Karen Howe did for a book called Empire of AI.
455. There was a recent article in The New Yorker by Ronan Farrow who documented a lot of this.
456. You might have slightly more generous words for Dennis Hassabis, who's the head of Google's DeepMind Project, and Dario Mode, who's the head of Anthropic.
457. But I think Elon and Sam are both people who have made many of those around them over the years feel truly did not have the best interests of not just humanity, but of anyone but themselves at heart.
458. What kind of response did you get to this wake call of yours Well there was a huge readership and I heard from a lot of people who really appreciated it I think the key thing that I had focused on which other readers felt gratified to see
459. was organizing all of these anxieties under the rubric of oligarchy and plutocracy and inequality.
460. And that previously, I think a lot of readers felt that they had some inchoate sense of unease about
461. how much of this was unfolding. They had worries or frustrations with how little was being done
462. at the political level to take control of it. But maybe they hadn't seen quite so clearly
463. the way that the AI story is unfolding really as an extension and intensification
464. of this longer-term story in which the very rich, especially the very rich coming out of Silicon
465. Valley, are exerting greater control over larger and larger shares of our public life.
466. And it connects to something I wrote a few months ago about the billionaire's tax proposed in California, in which I noted that just between the financial crisis, the Great Recession in 2008 and 2024, the amount of political donations given by American billionaires grew 150-fold from $16 million to $2.6 billion between 2010 and 2024.
467. Now, there are a lot of other ways in which our culture and our society and our political
468. economy have been distorted by money.
469. The fact that Elon bought Twitter and recalibrated its algorithms to benefit his own ideological
470. preferences has had actually an underappreciated effect on political discourse in this country.
471. But just at the level of sheer direct spending on elections, the influence of the very rich
472. has grown astronomically in this period.
473. And I think it's one of the main reasons why so many Americans feel so unsettled and so
474. so disempowered and so abandoned by our politics. They look up and see the world being run by a
475. small cohort of extremely rich people. And when they think about artificial intelligence,
476. making claims about the future of the country, the future of the economy, the future of society
477. around the world, they see the same pattern, an incredibly powerful, enriching technology that is
478. nevertheless, the work and owned by an extremely small number of people who are effectively imposing
479. their own preferences and worldviews on all the rest of us.
480. I've called it the rise of the corporate dictator. Your comments bring to mind one of the signs held
481. up in the artwork for your article reads as follows, not afraid of AI, afraid of billionaires
482. with AI, end quote, which reflects what you're saying. This is so fascinating, David. The great
483. thing about what you write is you don't mince words, but you don't engage in pie-in-the-sky
484. speculation as well. And you're absolutely right. The level of wealth compared to 30,
485. 40, 50 years in the hands of Jeff Bezos, Mark Zuckerberg, Sam Altman, Elon Musk especially,
486. is beyond anybody's speculation a few years ago.
487. We may see this year, Elon Musk become the first trillionaire
488. when SpaceX goes public and he owns 45% of the company.
489. So giving $250 million to Trump in his campaign by Elon Musk in 2024,
490. piece of change.
491. It's the piece of change.
492. Anyway, over to Hannah.
493. David, my question is about some of the naive, the super credulous people you've encountered as you cover AI.
494. I sure you talk to a lot of people where you like I don really think you believe what you saying You trying to push something You just trying to make money You don really believe that this will make anyone lives any better I would just look at the conversations and the statements that have been made by
495. the industry's leaders over the last couple of years. And what I mean by that is they just have
496. an understanding that the world is like a self-contained system in which a thought experiment
497. can go from 0.0 to 0.100 without any human complications along the way. And so you hear
498. these like offhanded promises like, well, AI is going to cure cancer or AI will make human labor
499. obsolete. And there doesn't seem to be any consideration of all of the intermediary steps
500. in which the technology needs to be embraced by humans, implemented by humans, serving human
501. customers in a political economy dominated by humans. These are not forces that can be swept
502. aside and replaced with a narrative where you just draw an upward sloping curve on a whiteboard and
503. say, by the time we get to this point, we know that there will be no place for humans in the
504. economy because humans are in charge of this world and will always be. So we may outsource
505. some amount of our decision-making to these tools, we may end up replacing a lot of employees with
506. these tools. But nevertheless, those will still be decisions that humans make and that they need
507. to be held accountable for. And again and again, when I'm talking to people in the AI world,
508. and especially when I listen to the sort of prognostications by their leaders,
509. I'm just struck by how little appreciation they have for the human messiness and conflict
510. that surrounds all kinds of cultural change, all kinds of economic change.
511. You know, one easy example is to go back to the political consequences of the China shock from a
512. couple of decades ago, where depending on how you measure, maybe we lost a couple of million
513. manufacturing jobs in the upper Midwest because of globalization, technology, and the entrance
514. of China into the WTO. We're a country of 350 million people. You could tell yourself on a
515. whiteboard in an economics seminar. Well, we'll be able to absorb those losses quite easily.
516. We'll grow faster, et cetera, et cetera. But the political consequences of that disruption
517. were extremely profound. We are still living through the earthquake that was unleashed by
518. those changes, even though they were relatively small. And almost nobody in AI seems to appreciate
519. that because they are so used to thinking of the world as something that unfolds in zeros and ones
520. on their computer screens and not as something that is contested by humans out in the real world.
521. And that's why I found that comment by Dario Amode, I was mentioning earlier, so striking.
522. He is the person who is the most humanistic, the most liberal of all of these leaders.
523. And he was saying that ideology would be retired by technology. That is to say,
524. he was imagining a future in which there would not be humans arranged in different
525. coalitional groups competing for power and status, but in which the question of where goods were
526. distributed, where wealth was distributed, and where resources were distributed would be managed
527. by a kind of silicon god, and all of us would defer to that judgment. And I guess we could see
528. a much crazier sci-fi future than I can imagine, but living in the world that I live in and living
529. in the world that I've lived in for the last four decades, it is just insane to me to think that
530. anything could come around that would retire human conflict and ideological contestation.
531. I think we heading if anything for a period of much more intense contestation Dr Yeah Actually that was an astonishing statement by him no more ideology I think what he meant it also no more ethical legal frameworks to restrain burgeoning technological risks
532. Anyway, we're out of time.
533. David Wallace-Wells, author of the recent article, The AI Backlash Will Be Ugly, and
534. No One Is Ready, Not Even Tech CEOs.
535. It was in the New York Times Sunday Magazine, May 17th, 2026.
536. For those of you who want to download it and use it as a discussion starter in your
537. neighborhoods as you elaborate your role of civic self-respect and recover our democratic
538. institutions that are now being dismantled. Thank you very much, David.
539. Thanks for having me. Great to talk to all of you.
540. We've been speaking with David Wallace-Wells. We will link to his work at the New York Times
541. at ralphnaderradiohour.com. I want to thank our guests again, Tyson Slocum and David Wallace-Wells.
542. For those of you listening on the radio, that's our show. For you podcast listeners,
543. stay tuned for some bonus material we call The Wrap-Up, featuring Francesco DeSantis with,
544. in case you haven't heard.
545. A transcript of this program will appear on the Ralph Nader Radio Hour
546. Substack site soon after the episode is posted.
547. The producers of the Ralph Nader Radio Hour are Jimmy Lee Wirt,
548. Hannah Feldman, and Matthew Maron.
549. Our executive producer is Alan Minsky.
550. Our theme music, Stand Up, Rise Up, was written and performed by Kemp Harris.
551. Our proofreader is Elizabeth Solomon.
552. Join us next week on the Ralph Nader Radio Hour.
553. Thank you, Ralph.
554. Thank you, everybody.
555. Step up, rise up, rise up
556. I know you wanna rise up
557. Stand up, stand up
558. You've been sitting way too long
559. Stand up, I think you'll step up
560. Step up, I think that beat will step up
561. Rise up, don't let the system hold you down
562. Stand up, stand up
563. You've been sitting way too long
564. Stand up, rise up
565. Step up, step up, I think you should step up
566. Rise up, you rise up
567. This is Chuck Foster.
568. I host Reggae Central every Sunday afternoon
569. right here on KPFK at 2 o'clock
570. where you'll hear ska, rocksteady, roots, dub and dancehall
571. every Sunday afternoon at 2
572. right here on KPFK 90.7 FM in LA.
573. It's one thing to love public radio,
574. but it's a special thing to support it.
575. Consider this.
576. If you've got a car you don't need anymore
577. or you've got one that is simply too expensive to repair,
578. arrange to donate it to KPFK.
579. It's easy for you, pickup is free, and it could be worth hundreds of dollars in support.
580. Thanks in advance, and here is how to get started.
581. Go to kpfk.org slash cars or call 877-573-5288.
582. That is 1-877-KPFKAUTO.
583. and learn more.
584. Hi, this is Curt Ellen. You're listening to KPFK.