Click to subscribe on: Apple / Spotify / Amazon / YouTube / Rumble
Robert Scheer is joined by investigative journalist Peter Byrne on this episode of Scheer Intelligence to discuss the state of artificial intelligence and its omnipresence in both civilian and military life. Byrne dives into his “Military AI Watch” series for Project Censored, where he explores the militarization of AI and the for-profit development by the hands of Silicon Valley giants and the American military and intelligence apparatus.
Byrne explains the motivations by some of the key actors involved in the industry including Palantir co-founder Peter Thiel, Anduril founder Palmer Luckey and others. The ties to the American military are unavoidable and they are crucial to understanding the backbone of the AI industry. “The United States military and the intelligence agencies are actively engaged in creating what is called a Joint Awareness Domain Command and Control System by which they can surveil the entire global battle space and operate wars from your computers in the Pentagon,” Byrne tells Scheer.
“Artificial intelligence is key to that.”
Credits
Host:
Executive Producer:
Video Producer:
Introduction:
Transcript
This transcript was produced by an automated transcription service. Please refer to the audio interview to ensure accuracy.
Robert Scheer
Hi, this is Robert Scheer with another edition of Scheer Intelligence where the intelligence comes to my guess. In this case, a well-known journalist written a lot about national security issues, Peter Byrne, and he’s done a project or is doing a project for Project Censored, which is probably, I don’t know, one of the most useful organizations out there telling us what we don’t get to hear and read in the media. And this series he’s already done two of the 10 really is trying to assess the impact of artificial intelligence on our lives, our security, misuse by the military and what have you. I’m going to let Peter take it from there and tell us what you’re doing.
Peter Byrne
Well, it’s a 10 part series that’ll run once a month through January and is sponsored by Project Censored, which I’ve worked with in the past. It’s a great organization. The first article in the series, which is called Military AI Watch, is called “One Ring to Rule Them All.” It’s a 7,000 word deep dive into the nature of Palantir and Anduril in terms of antitrust issues, in terms of their history, their fiscal history, their non-payment of federal taxes, things like that.
And it also gets into the technology and the science of how the programs that they’re selling, the AI programs and devices and weapons that they’re selling to the Pentagon don’t work as advertised and what the pitfalls are. The second story in the series is about Stargate. It’s about the data centers that are proliferating around the world. And it focuses on how World War III or maybe it’s World War IV is likely to be fought inside the data centers, which are highly polluting, energy-intensive disasters.
And it also goes into the financial aspect about how Middle Eastern sovereign funds and connected to the Trumps are investing in this. It’s quite the boondoggle.
Robert Scheer
All right, so let’s take it to Palantir now and the first article. And basically we’re talking about, you know, this is no longer just a private sector. We’re talking about a company that basically had the CIA as an investor. Obviously it had a big private investment, but it’s only client for the first three years.
And I wrote about this in my own book, “They Know Everything About You,” was the CIA. And they got to play with the CIA’s data and use it to develop their models, which lead to this kind of use of artificial intelligence, supercomputers using a lot of power, assimilating a lot of data and drawing conclusions from it. It even led to predictive policing in Los Angeles, but it’s been worldwide in its impact.
So why don’t we begin there. Palantir started with money from a CIA organization called In-Q-Tel and they pumped in some, but that wasn’t really the whole point, is that they’ve been a collaborator with the CIA from the beginning. So take it from there.
Peter Byrne
Collaborated with the CIA and the National Security Agency and Homeland Security, AKA ICE. Most of their contracts, more than half their contracts worldwide are with military and intelligence agencies, as well as with huge multinationals like Coca-Cola. But recently, Palantir went public in 2020. so now we have a glimpse into its interior workings via the Securities and Exchange Commission.
It’s quite interesting. They’ve never paid any federal taxes and they recently allocated $5 billion to stock paybacks, which after Trump was elected, they exercised. So the Palantir board, Peter Thiel and Alex Karp and a Wall Street Journal reporter, former Wall Street Journal reporter and other cheerleaders of the weaponization of AI got really rich in February.
Do we want to talk about the trope that Thiel names his companies after objects in Tolkien’s Lord of the Rings, which is kind of interesting thing that I get into in the start of the deep dive, One Ring to Rule Them All. The short story is that Palantir is named after the Seeing Stone in the Lord of the Rings, where the Dark Lord Soren can see where all of his enemies are and then go to annihilate them.
He’s got Mithril Capital, which is a Tolkien word. His apprentice, J.D. Vance, who apparently now has high office in the United States, has a company named after a Tolkien trope called Narya Capital. So you’ve got these Silicon Valley oligarchs who depend on this kind of ancient mantra that came out of Sausalito back in the early 90s, that libertarians and hippies invented the internet. Actually, the CIA and the military pretty much has dominated the internet since the very beginning, as say, Yasha Levine showed in his great book, 2018, Surveillance Valley, Google was created by the Department of Defense and the CIA. And that continues.
So what I’m looking at throughout this series is the financial and the scientific and technical aspects of the militarization of artificial intelligence. The third article, which I’m working on right now, is about how the electromagnetic spectrum is being weaponized. We talk about 5G, for example, with our cell phones. Most of us don’t know that 5G is a method of connecting what is called the internet of military things, which includes your smartwatch and your iPad and your phone and the surveillance cameras on your neighborhood cell phone tower, which is basically a vast global network that due to 5G technology is being turned into a surveillance mechanism that’s also capable of jamming drones and engaging in battle space activities.
So that’s the next article going on about what 5G actually means for the military. And we’ll be following up over the next eight months or so with deep dives into how Eric Schmidt and Peter Thiel and others are using their nonprofits to launder, as it were, quantities of stock and control markets.
As you can see from my office here, I’ve gathered a vast amount of material over the past several years on this. And only a fraction of it’s going to make it into the 10 articles, but we’ll see where we go from there.
Robert Scheer
So what’s the basic theme? mean, because artificial intelligence is being presented as a boon, not to working people who quite often are threatened with the loss of their jobs and making their skill set seem meager by comparison to what huge computers can do. But what you’re really painting is a picture of world politics, world governance.
War and peace and the survival of the planet, the survival of humanity, being in the hands of the people who can use artificial intelligence and its enormously expensive apparatus to basically control reality.
Peter Byrne
Yeah, if you want to look at the one version of the future, look at the Xinjiang province in China, where the Uyghur people, two million strong, they’re a minority in China, so-called. They’re not Han. And for the last 20 years, their entire society has been surveilled and controlled through various iterations of artificial intelligence. The latest iteration, is large language models, is simply that. It’s the most modern addition, the most modern technology of artificial intelligence, which is a much broader concept.
At any rate, Apple and Google and the standard, you know, Meta and the rest of them are all actually collaborating with the Chinese government to conduct this like social experiment in Xinjiang province where their cameras are actually on 24 hours a day in people’s homes. And if you say the wrong thing, you’re detained and put into forced labor camps. People have been writing books about this. It’s not trivial. And it’s done completely with the technology made available by Silicon Valley, know, bohemus. So, you know, what was it Leonard Cohen had his song about? I’ve seen the future and it is where they want to control every living soul.
That is the ruling class’s aim is to have complete control. I mean, Trump and the rest of them, they’re talking about re-industrializing America with data centers, are not talking about bringing good jobs here. They’re talking about, as Commerce Secretary Lutnick said, millions and millions of people screwing in little screws into iPhones. What they want to do is they want to bring essentially slave labor back to the United States.
Robert Scheer
So let’s stick to the theme of your first article, because we’re going to be doing this as you reveal more of this work. What is the takeaway?
Peter Byrne
The takeaway is that in line with what I was just talking about, the United States military and the intelligence agencies are actively engaged in creating what is called a Joint Awareness Domain Command and Control System by which they can surveil the entire global battle space and operate wars from your computers in the Pentagon. And artificial intelligence is key to that. As my article shows, it doesn’t work. It doesn’t work for a number of reasons, including that it’s too complicated to actually unite into one picture of what’s going on.
And one of the reasons it’s too complicated is because it’s really easy to jam electronic signals. It’s easy to insert and spoof false data, poisoning the data, so-called. The idea of having a surveillable, complete global battle space is not feasible for that reason. But there’s another reason why it’s not feasible is because the Army, the Navy, and the Air Force and the intelligence agencies all have their own versions of this joint awareness command and control system and they don’t talk to each other.
It’s just typical Pentagon contracting as usual. At this time, the organizations like Palantir and Anduril, also controlled by Peter Thiel, are competing with Lockheed and Raytheon and the big primes to basically reform the American military apparatus in terms of being operated through artificial intelligence and autonomous weapons.
Robert Scheer
And what’s the downside of that? If it doesn’t work, you know, that’s just a waste of money. But when you talk about surveillance, when you talk about basically talking about excluding the public from participating in deciding when to go to war and how to conduct it.
Peter Byrne
Well, it doesn’t work in the sense that it doesn’t work as they envision it. It will not give them complete control because it breaks down in so many areas. But where it’s particularly dangerous is that, for example, the nuclear command and control system, the so-called NC3, is also tied into it. And as we know, we’ve almost had huge nuclear armageddons due to computer glitches, like five or six times at least that we know about in the past.
So the more complicated a system becomes, the more fragile it is, the easier it is for mistakes to proliferate inside of it. So we could end up with wars leading up to nuclear wars, which occur literally in nanoseconds, because the military is increasingly offloading its decision-making capabilities to artificial intelligence.
Primarily in the form of these large language models these chats which are are prone to hallucinations I mean at least 20% of the data that you get from your local chat is is gonna be wrong so…
Robert Scheer
But again, I mean, I’m trying to get at why the alarm and what does it do to democracy? What does it do to developing a saner world? And yes, there’s this literary reference in it to power and so forth. But is this a fantasy? You mentioned Peter Thiel. We haven’t really gone into his role, but he made a lot of money with PayPal. He’s one of the people who founded Palantir, but what is the scheme here? Are you suggesting that they’re going to be able to use this system to control our lives in ways we can’t imagine, to control human history in ways we can hardly imagine? What’s up?
Peter Byrne
The head of Anduril, man named Palmer Lucky, he famously stated recently that the way that things are going with artificial intelligence and warfare, pretty soon adversaries will just be able to look at their enemies capabilities and they’ll surrender. So they won’t have to go to war. It’ll all just be comparing weaponized AI driven technologies and rational forces will decide that they can’t beat each other. That’s not the way war works. War isn’t a game. War is not a rational game theory type of driven exercise.
It’s full of fog and mistakes. And once it gets started, it’s really hard to stop. And war is also about killing people. It’s not about computers competing with each other. So, you know, as we’ve seen with so-called precision strikes in Afghanistan and Iraq and elsewhere, Somalia, Gaza, which is heavily influenced by Palantir and Anduril and Microsoft and Google AI-driven programs killing the Palestinian people in this massive genocide, the collateral damage in all of this is extraordinary. And when you throw in the nuclear threat, it’s existential.
Robert Scheer
And so, what is the, again, I want to push back here. The danger is that these people will think that they can make rational decisions. I mean, if we blow it all up, obviously they are not impervious to it. There are plans to live on Mars or somewhere, but that’s probably not going to be realized even though they think they can extend their own lives. I just want to get a picture because we got this category that you’re dealing with in this ten part series of artificial intelligence. It assumes a higher level of intelligence than humans are capable of. The argument then is there an Orwellian dimension here that it just puts everything out of control?
Peter Byrne
gonna push back on what you said. Artificial intelligence does not assume a higher level of intelligence. That’s just marketing hype. Artificial intelligence isn’t intelligent at all. It doesn’t resemble the brain or the thinking process or how decisions are made in human societies at all. It is simply a glorified search engine, which basically I explained in great detail in these articles. So the danger is that people actually believe that it is intelligent and…
There’s a certain type of bias that people are recognizing, which is to give deference to artificial intelligence, especially in military circles, so that the idea the Pentagon says is we’ll always have a human in the kill chain, a human in the loop to decide whether not to pull the trigger. But what happens is that the AI will say, well, we have a suspicious character in Qatar going down the street with what could be a load of bombs towards the American embassy should we annihilate it.
And in theory, like some, you know, second lieutenant in Creech, Nevada will go, I don’t think we should, it might be babies. But no, this is not what’s gonna be happening. There will be no human in the loop. The AI system will simply be able to play the odds and say, well, there’s a chance that it’s an enemy. And so I’m going to…
I, being the machine, going to authorize the strike. Now, you know, the humans make plenty of mistakes too. So the problem is not so much the existence of AI, it’s how we’re using it. And we’re using it to perpetuate an imperial military war structure, which is destroying people, you know, by blowing them up and controlling their societies, et cetera, et cetera, throughout the world.
So the problem as far as I’m concerned is actually not AI, it’s us.
Robert Scheer
That is worth the price of admission here to what we’re trying to do. I mean, I’m glad you pushed back. that is really as we should mention how people get to read the first two articles. I’ll post them alongside our interview here, our podcast. But how else? I mean, because we’re going to run out of time on this first discussion. But the best thing, of course, is to read the articles. How do they get them? Aside from by going…
Peter Byrne
You can just go to the Project Censored website and it’ll be there. you can actually, I hate to say, I’m not going to say Google, but you can put into your favorite search engine, Military AI Watch, and my name or Project Censored, my name is Byrne, B-Y-R-N-E, and it’ll pop right up. You might have to navigate through a few actual Military AI Watches, you know, which they’re going to have on there and try to sell you. But if you, concentrated on Project Censored and military AI watch, you would take into the existence of the articles on the computer.
Robert Scheer
Yeah, and we’ll post them on with this first, which I hope will be first of a whole series that we do on this. And I love your pushback. That’s what I’m here to do, to be educated by you and help our readers get educated by you. So keep pushing back. And I want to explore, and we only have another 10, nine minutes here. I want to explore this idea that artificial intelligence is not intelligent.
It’s being mystified for people to bow down to it. you know, even students in school give me a term paper. It’s obviously better than I can do. I’ll get a good grade and so forth. And there are manifestations of it working. You know, I’ve even tried putting our own transcript here through versions of AI and see what they come up with, you know, for an introduction, just, you know, full disclosure. But clearly, to have any worth, have to, least at this stage of the game, go over it and change it and manipulate it and maybe even use it as a teaching tool in some sense. But I know in your article you stress that there’s a lot of hype about this. So let’s, in the minutes we have left, tell us about the hype.
Peter Byrne
Like I said, it’s a glorified search engine. There’s plenty of great science papers that make that same argument, although maybe not quite so bluntly, although some do. People have to understand that the large language models are not replicas of the brains. They’re not neural in that sense. What they are is a collection of what are called perceptrons, which are basically mathematical functions that are encoded into chips, as it were. And the perceptrons are fed trillions and trillions of data points, information that is training the perceptrons on how to make correlations so that it can make predictions between pieces of words or pieces of images.
So if you’ve got a cat’s foot and a cat’s head, it will be able to give you the image of an entire cat by putting that together and then guessing what comes in between. Now, in order to do that, it’s got to have millions and millions of pictures of cats, right? Now, the training corpus of the large language models is the internet. It’s Wikipedia, it’s Reddit, it’s the common crawl.
It is not scientific databases, it’s not governmental databases. Most of the knowledge bases that are scientifically based in our societies are not allowing the bots to crawl from Chat GPT, open AI and the rest of them, entropic. So consequently, the training data that is being used to feed these perceptrons is largely internet garbage is full of misconceptions and it is no surprise consequently that when you ask Chat GPT a question, least 20% of the time it’s going to come up with a complete non-secretory.
Like recently Elon Musk’s Grok has been saying that, we have to stop the genocide of white people in South Africa in response to just completely non-secretory types of questions.
So the problem is that this is incurable. I can point people to any number of really good scientific technical papers delineating why it is impossible to make large language models tell the truth, as it were, or even to make cogent arguments. Now, if they’re trained on very narrow data sets, like, say, x-rays of mammograms or something like that, they’re prone to make less mistakes.
But this whole idea that you can ask a generalized question of a chat leads to the type of confusions that we have now. It also uses vast amounts of energy, so much that between the energy use and the intellectual confusions it’s leading to, I think it was Meta the other day asked people to stop being polite to the chats because it was using up too much energy.
So, you know, people are psychologically venerating the chats. It’s almost as if they picked up a stone and it started talking to them. They believe that it actually has human agency or that it can replicate it. It doesn’t. It is basically just predicting whether pieces of words or pieces of images can be correlated to each other in this vast phase space that is filled up with pieces of words and images that have mathematical proximity towards each other and therefore a chance of making some kind of sense to humans when the program sifts through trillions and trillions of possible correlations to come up with the one that thinks it will answer the question you asked it. Now as far as using copyle and stuff like that, I read a lot of like multi hundred page audits and RAND reports and government military documents and things like that. And I’ve experimented with seeing if they can sum it up and they don’t.
They basically just make the most superficial summary of what’s in the document. They don’t do what an investigative reporter needs to do, which is to find the contradictions and the lies, to find, you know, the hidden meanings behind the hype.
So when I talk about hype, I’m gonna flog right now a new book by Emily Bender and her colleagues. It’s called The AI Con. It just came out. And Emily Bender famously wrote a paper about four years ago…
Robert Scheer
Well, we’re going to end this session, but I want to get back with you because you have a clarity and a skepticism about this that is really very refreshing. I’m going to post the articles that you’ve written so far, and we’re going to do this every few weeks.
All right, well, that’s going to be the next conversation. We can do it next week if you want. Let’s aim for that. Why not catch up a couple of weeks? OK. All right. But it really is a cautionary tale so far that we can’t ignore. So let me just thank my executive producer, Joshua Scheer, for putting this all together, and Diego Ramos, who writes the introduction so we don’t use AI.
And Max Jones who does the video and I want to thank the JKW Foundation for giving us support. In memory of Jean Stein, late Jean Stein, a great writer who exposed a lot of wrongdoing and Integrity Media and that has supported us in effort to have a better press as they do other groups. So see you next week with another edition of this Scheer Intelligence.
Please share this story and help us grow our network!
Robert Scheer
Robert Scheer, publisher of ScheerPost and award-winning journalist and author of a dozen books, has a reputation for strong social and political writing over his nearly 60 years as a journalist. His award-winning journalism has appeared in publications nationwide—he was Vietnam correspondent and editor of Ramparts magazine, national correspondent and columnist for the Los Angeles Times—and his in-depth interviews with Jimmy Carter, Richard Nixon, Ronald Reagan, Bill Clinton, Mikhail Gorbachev and others made headlines. He co-hosted KCRW’s political program Left, Right and Center and now hosts Scheer Intelligence, an independent ScheerPost podcast with people who discuss the day’s most important issues.
Editor’s Note: At a moment when the once vaunted model of responsible journalism is overwhelmingly the play thing of self-serving billionaires and their corporate scribes, alternatives of integrity are desperately needed, and ScheerPost is one of them. Please support our independent journalism by contributing to our online donation platform, Network for Good, or send a check to our new PO Box. We can’t thank you enough, and promise to keep bringing you this kind of vital news.
You can also make a donation to our PayPal or subscribe to our Patreon.