Transcription
Palantir here has been called the world's most evil company more than once, and there are a number of factors that have contributed to that. It's been accused of lending its technology to the Israeli military's Lamar AI targeting system, which has marked tens of thousands of Palestinians in Gaza for assassination by drone, bomb, and missile.
Likewise, Palantir technology has also been alleged to be used by the Gospel, another targeting system that has allowed the Israeli military to go from creating 50 targets a year to 100 per day. Now, Palantir has denied that its software is used by those two tools. Though in a letter to Amnesty International, after they queried these allegations, the US tech company did say it was "proud to support Israeli Defense and National Security missions in other programs and contexts." And in early 2024, just 3 months into Israel's genocidal assault on Gaza, Palantir signed a contract to provide battle technology for war-related missions.
So what it doesn't deny, though, is that its software has been used by the US military in Iran. This was Palantir Chief Technological Officer Shyam Zankar speaking on Bloomberg last month.
"People will reflect back and say this is the the first large-scale combat operation that was um really uh driven, enhanced, made substantially more productive with technology, with AI. Uh, and you know, the word 'targeting,' I think to the to the lay public has some sort of connotation. and it sounds like someone's shooting, but it's actually a very bureaucratic, large-scale doctrinal process. So if you thought about that, like any, you know, I have a value chain to manage, I have a production process to manage, I have a targeting process to manage. Um, being able to accelerate that. So what in, if you think about Gulf War I, uh, I think, or Gulf War I, sorry, Gulf War II, uh, we did about a thousand targets. It was 6 months of planning for roughly 50 to 100 people. And in this, in this conflict, you're looking at that equivalent of work for twice as many targets was done by one person in two weeks."
It's really interesting that Palantir, uh, obviously sells a lot of its technology or or is involved in contracts with the military, but it's also involved in really mundane things like delivering like like frozen chicken supply supply chains for a very big, um, uh, you know, chicken abattoir, basically, or chicken meat product supplier in in the US. And the way that he talks there is really to just say, like, actually, those two processes are just exactly the same. Uh, shipping frozen chicken nuggets out to where they're most needed, depending on how on people's movements and things in different parts of the country, is just the same, um, as targeting people for death, uh, through the military. And the outcome that you want is like increased efficiencies in both. I find that like a deeply extraordinary and corrupt way of of speaking about two vastly different kinds of, um, kinds of activity. Um, and it is what they produce: extraordinarily efficient mass murder.
According to the New Yorker, Maven allows military officials to select up to 80 targets an hour, dispatching them in just four clicks. And Shankar is right. The war in Iran is the first AI war, but it certainly won't be the last, with both the US and Israel said to be using their illegal assaults in Iran as a test of its autonomous tech. At the heart of it is Maven, a CIA-backed project to create a lightning-fast AI-assisted targeting system for the US military. Originally handed to Google, the company dropped the project after 3,000 staff members revolted at the idea. Since 2018, the $1.3 billion operational contract has been held by Palantir instead.
Sankh went on to justify Palantir's involvement.
"This is a realization of the third offset strategy. You know, the first offset was nuclear weapons. Second offset was precision-guided munitions. How do we? We're not going to, we're not going to, we don't want to engage symmetrically with the Soviets like every missile they have, we need to have a missile. No, we're going to have more precise missiles. So we'll need less of them as a consequence. Don't forget in World War II, only 6% of bombs dropped hit."
"This is very asymmetric, right now."
"And and and that is exactly the point. That is deterrence."
"It may be a deterrent, but a weapon is only a deterrent if it's used at least once. Um, and this one isn't always accurate either. On the first day of the US and Israel strikes in Iran, one of the 10,000 targets hit turned out to be a school in the middle of morning lessons. That resulted in the killing of more than 180 people, the vast majority of them schoolgirls under the age of 13. Questions have been asked in the US Congress as well as in the media about the role of Maven in that atrocity. So far, few have been answered, though."
He, head of Palantir's UK operations, Louis Mosley, recently explained to the BBC why mistaken or even deliberate attacks on incorrect targets could never be Palantir's fault.
"If an AI system recommends a target that later turns out to involve civilians, where does responsibility for that lie? That is with the military organization."
"Okay. And critics warn that AI systems can encourage faster strikes, you know, and that they're they're making the time from, uh, planning a mission to executing that mission much faster. But doesn't that introduce quite a bit of risk in terms of additional things going wrong? By scaling this activity and moving at such pace and such speed, doesn't that introduce additional risk in terms of stuff going wrong?"
"So it, it, that the increase in speed is is a consequence of the increased efficiency, which comes from the greater accuracy with which those decisions can be taken at speed. So it's not that the the systems are just moving faster without the accuracy also keeping up. It's obviously not always very accurate because, uh, a school full of school girls was was hit, was hit. I mean, I think the thing about speed as well is that, you know, like it goes so fast that it defies human checking, the possibility of of oversight. It's just the system overseeing it itself, and and human beings just sort of delivering the the targets, uh, it's been given."
This weekend, Palantir did seem to answer a different question, though. Are you an out and proud fascist company? The answer appears to have been yes. In a lengthy post on X, running to over a thousand words, Palantir put forward what looks very much like a 22-point manifesto. Don't worry, I am not reading all of that. It's in fact a summary, kind of, of this book, *The Technological Republic: Hard Power, Soft Belief, and the Future of the West*, written by Palantir co-founder Alex Karp last year.
Now, here are a few of those 22 points that, uh, Palantir put out:
* Silicon Valley owes a moral debt to the country that made its rise possible.
* The engineering elite of Silicon Valley has an affirmative obligation to participate in the defense of the nation.
* The limits of soft power of soaring rhetoric alone have been exposed.
* The ability of free and democratic societies to prevail requires something more than moral appeal. It requires hard power, and hard power in this century will be built on software.
* The question is not whether weapons will be AI weapons will be built. It's who will build them and for what purpose.
* Our adversaries will not pause to indulge in theatrical debates, theatrical debates about the merits of developing technologies with critical military and national security applications. They will proceed.
* National service should be a universal duty. We should, as a society, seriously consider moving away from an all-volunteer force and only fight the next war if everyone shares in the risk and the cost.
* The atomic age is ending. One age of deterrence, the atomic age is ending, and a new era of deterrence built on AI is set to begin.
And these are just a few shorter bullet points without the elaboration from that list:
* No other country in the history of the world has advanced progressive values more than this one. That means the US.
* American power has made possible an extraordinarily long peace.
* Silicon Valley must play a role in addressing violent crime.
* Some cultures have produced vital advances. Others remain dysfunctional and regressive.
* We must resist the shallow temptation of a vacant and hollow pluralism.
So you have a whole bunch of things there. Um, Harriet, you have, um, you know, the idea that it's a civic duty of a company like Palantir to tackle crime, for example, and we know that they have been, you know, use their technology and all kinds of, um, pre-crime kind of identification stuff. It's a civic duty of of people to be conscripted into the military, uh, so that when we go to wars, people share the cost. I mean, people already share the cost, right? They pay for it. Uh, you've got the idea that, um, that Palantir has a civic duty, uh, to defend the state, which invariably means creating huge offensive weapons and systems, which, you know, aren't about like an Iron Dome or something, which are about killing people in other countries. Um, there are all the elements of a sort of very frightening kind of picture of the relationship between, um, software company and the state. I mean, the other thing is, of course, is pitching AI AI weapons as the new deterrent. So, um, taking over where the nuclear bomb leaves off as the the main way of maintaining peace, which is that, you know, you sort of show people that you can take them out in a really, really hard way if they, if they, if they mess with you. So it's a kind of might-is-right mentality.
You've done a lot of writing on Palantir. Um, does any of this surprise you?
"Not in the slightest. Not at all. No part of this surprises me. Um, this is a really disturbing, authoritarian, technofascist manifesto. It's pitching for this highly militarized state, um, which would obviously benefit Palantir's bottom line. And as we know that under the Trump administration, um, Palantir has an absolute wealth of government contracts, and it has been instrumental in facilitating Trump's, um, immigration enforcement crackdown via ICE, and that Palantir has developed and created tools specifically for ICE, including ones that, um, actually use people's healthcare data to populate maps to show where the most people who might be targeted for ICE raids are. I mean, this is even, even before we go on to Palantir's, um, involvement in Israel's genocide in Gaza, this is deeply, deeply disturbing, and an absolute shame upon our government for allowing this company, this really dangerous company, which I think should actually be treated as a hostile foreign actor, into the most sensitive and privileged parts of the British state, including things like nuclear weapons, the Ministry of Defense, and most egregious of all, the NHS. Our private healthcare data can be accessed by this company. I mean, the guardrails in place by NHS England do not stop Palantir employees for being able to access identifiable, um, healthcare data from the British public. Like, that is absolutely true. And shame, actually, also on Politico for allowing their newsletter to be, um, sponsored by Palantir all of last week. I was absolutely disgusted to see that."
"This week as well, because I read it this morning and it's still great. Great."
"Yeah. I mean, the deepening NHS access by Palantir. We've got crossbench MPs now calling for that break clause in February next year to be triggered to get us out of this contract with this really dangerous US spytech firm. Um, people, uh, MPs from Labour, from the Lib Dems, and obviously Green Party leader Zack Plansky has been pushing for this as well. We've got a system that's now embedded in the NHS which runs on Foundry, which is the civil equivalent of Gotham, which is the Palantir tech. It's their software platform that is used by militaries around the world. Um, we've got Palantir staff and other government departments in this country being able to access our healthcare data, which seems like worryingly familiar. When you look at what ICE can do with healthcare data, like Obamacare data in the US, like that's really, really scary. And we've got NHS provider trusts getting repeated requests every week for people's healthcare data from Palantir. And that is information we've got via, uh, campaign groups like Medact, which include, which are made up of NHS staff. This is utterly terrifying. Um, we know that they got in the door through COVID, through COVID databases, and like it was part of their strategy to pitch low to go to like gobble up other businesses, um, and to sort of like flatten out any criticism that way. And this COVID database, it paved the way for way more lucrative contracts within the NHS. Um, and people are warning now that our NHS may become dependent on Palantir. It's going to become a monopoly supplier. It's called like vendor lock-in, that's what they call it. Um, and we haven't got any, like, the NHS hasn't got proprietary anything over this system. The federated data platform is Palantir software. It absolutely owns it. Um, I've, I see, like, different things. So, the the FT this week was like, well, maybe it actually saves money. They, they did a, they published like a positive editorial leader on it. How can that be the absolute, um, standard for why we should let this company into our NHS and have access to our data? A company of which its co-founder and chairperson Peter Thiel has said that the NHS essentially needs ripping up, and that our love for the NHS as British people is Stockholm syndrome. Like, how has this been allowed to happen?"
"How has it been allowed to happen? And I mean, like for me, it's just, you know, Louis Mosley was on, like Politics Live last week, and he was saying, uh, Faisal Shaheen asked about, you know, the use of of the software for, you know, war, and he was like, well, that's like saying, you know, like Microsoft, it's, you know, Word or whatever PowerPoint is used in, you know, these applications, but it's just, I just don't think it's the same at all, right? Because the whole point of AI is it's supposed to develop a kind of, I mean, not that Palantir is like AI in one sense of the word, but this kind of algorithmic software is, it's supposed to develop a kind of dependence, uh, which has like applications in every single part of, or most parts of, uh, whatever kind of op, you know, operations or, um, exercises are carried out within the NHS. So even though, you know, maybe it's just used in a limited way at one, you know, at this stage, the idea is that it will grow and spread everywhere. And what we're talking about growing and spreading everywhere is the very same, is a product of the very same company which is creating, has been explicit in wanting to create the best AI soft AI weapons out there in order to act as a deterrent effect for the US. And as I said in the piece, you know, like nothing, it's not a deterrent unless you use it once, so people can see what it does. I think that if, if we're not taking this seriously, we're really not paying attention."
"Sleep at the wheel."
"Yeah, absolutely. It's this company has already been allowed to get its tentacles into the very heart of the British government. Like, I, I feel like there has been such failure up until this point that now the alarm bells are ringing too loud for anybody not to like know about this and have like a really, really deep, visceral reaction to it. Um, I referenced that FT piece earlier. They said that, um, people's criticisms of Palantir is usually emotional."
"Louis Mosley said on on Public Live said it was ideological."
"Oh, okay. Guess."
"Nothing wrong with an ideology, though."