Daniel sent us this one — a topic that's been floating around the edges of news coverage for two decades now but almost never gets dissected clearly. He's asking about black sites. What are they, exactly — is that a real intelligence term or just something the Bourne movies invented? Which ones have actually been confirmed by courts and investigations versus which ones live in rumor territory? And what's the strategic and legal logic behind secret detention in the first place? He wants us to separate the documented from the speculative and give people a framework for evaluating claims when they hear them.
This isn't ancient history either. Just this week there's been movement on the Al-Nashiri case against Romania at the European Court of Human Rights — still pending, still unresolved. The legal fallout from the CIA's rendition and detention program is actively shaping international law right now, not twenty years ago.
That's the thing that jumps out when you actually dig into this — the gap between the pop-culture image and the legal paper trail. People picture some Bond-villain lair, but the reality is a network of facilities that have been litigated in actual courtrooms, with actual rulings, compensation payments, the whole machinery of law grinding through it.
Which is exactly why we need to start with definitions. Because the term black site — it's not what most people think it is.
It was never an official CIA classification. Never appeared in any agency manual or formal doctrine. It emerged as slang inside the intelligence community — shorthand for a facility that was unacknowledged, off the books, not declared to the host government's own security services in some cases. Then the Washington Post got hold of it in two thousand five, Dana Priest's reporting blew the lid off, and suddenly it's everywhere. By the time the Senate Intelligence Committee dropped its torture report in twenty fourteen, black site had become the catch-all term for the entire secret detention archipelago.
That's where the definitional sloppiness creates real confusion. Because not all black sites were the same thing. I think of it as a three-tier taxonomy. Tier one — the CIA black sites proper, the facilities run under the Rendition, Detention, and Interrogation program for high-value detainees. These are the ones where enhanced interrogation techniques were used. Tier two — military black sites, places like the Bagram Theater Internment Facility or Camp Bucca, which were secret or semi-secret but operated under different legal authorities. And tier three — proxy facilities, where the U.essentially outsourced detention to another country's intelligence service.
The "we'll hold them, you ask the questions" model.
Syria's Air Force Intelligence prison was a notorious example of that third tier. would render detainees to Syrian custody knowing full well what Syrian interrogation looked like.
Here's another distinction that gets lost — black site is not synonymous with torture center. Some of these facilities were transit hubs. Diego Garcia, the British island in the Indian Ocean, was used as a refueling and transfer point for rendition flights. The UK government confirmed that in two thousand eight. Nobody was waterboarded on Diego Garcia. It was infrastructure, a logistics node in a secret network. Calling it a torture site is factually wrong, but calling it a black site — in the sense of an unacknowledged facility supporting the rendition program — that's accurate.
With that definitional ground cleared, let's get into the cases where we actually have evidence. And I want to emphasize — we're not talking about journalism here, or at least not only journalism. We're talking about court rulings. The kind of evidence that would be admissible in a courtroom.
Let's start with the best-documented one, because it's the case that really cracked the whole thing open.
It's a Polish intelligence training base in the northeastern part of the country, surrounded by forest, very secluded. In two thousand fourteen, the European Court of Human Rights issued a ruling — actually two rulings, consolidated — Al-Nashiri versus Poland and Abu Zubaydah versus Poland. The court found that Poland had allowed the CIA to operate a secret detention facility on its territory, that Abd al-Rahim al-Nashiri and Abu Zubaydah were held there in two thousand two to two thousand three, and that they were subjected to treatment that violated Article Three of the European Convention on Human Rights — the prohibition on torture and inhuman or degrading treatment.
Article Five — the right to liberty. They were held incommunicado, no access to lawyers, no contact with family, no judicial review of their detention.
The evidence chain on this one is remarkable. You had the Senate Intelligence Committee report, which described the facility in detail without naming Poland explicitly — but the geographic and operational details matched Stare Kiejkuty exactly. You had flight records showing CIA-chartered aircraft landing at Szymany airport, which is about twenty kilometers from the base. You had former CIA officers talking to journalists. And you had the Polish government's own internal investigation, which former President Aleksander Kwaśniewski later acknowledged had confirmed the site's existence.
Poland ended up paying compensation — a hundred thousand euros each to al-Nashiri and Zubaydah, plus covering their legal costs. And in twenty twenty-three, a Polish court convicted a former head of the intelligence agency for facilitating the site. That's not speculation. That's a criminal conviction.
The second confirmed site is Lithuania's Project Number One. That's what the CIA called it internally — Project Number One. It was a facility near Vilnius, a converted riding school on the outskirts of the capital. Lithuania's own parliament conducted an investigation in twenty eleven and confirmed that the site existed, that it was provided to the CIA, and that the Lithuanian government knew what it would be used for.
The ECHR ruled against Lithuania in twenty eighteen — Abu Zubaydah versus Lithuania. The court found that Lithuania had knowingly facilitated secret detention and had violated the same articles as Poland. The really damning detail was that Lithuania provided the facility in exchange for what was described as "support for its NATO accession aspirations." It was a geopolitical transaction.
Which we'll come back to when we talk about the strategic logic, because that exchange — hosting a black site in return for security guarantees — is a recurring pattern.
Then there's Romania. The facility was code-named Bright Light, located in Bucharest, and it was used for some of the highest-value detainees in the entire program — including Khalid Sheikh Mohammed, the architect of the nine-eleven attacks. The Council of Europe's Marty Report in two thousand six identified Romania as a host country, and the Senate report confirmed the details. The ECHR case, Al-Nashiri versus Romania, was filed in twenty fifteen and as of July twenty twenty-six, it's still pending.
Which is staggering when you think about it — eleven years and no ruling yet. That tells you something about the complexity of these cases and the reluctance of courts to wade into intelligence matters.
The first black site post-nine-eleven wasn't in Europe at all. It was in Thailand. Code-named Cat's Eye, located near Udon Thani in the northeast of the country, reportedly a former U.This is where Abu Zubaydah was first held after his capture in Pakistan in two thousand two. And this is where the waterboarding happened.
Eighty-three times in August two thousand two alone. That number comes directly from the Senate report. Eighty-three waterboarding sessions in a single month. The CIA's own cables from Cat's Eye described Zubaydah as being in a state of complete physical and psychological collapse, and yet they kept going. The Thai government has never officially acknowledged the facility's existence. There's been no investigation, no court case, no parliamentary inquiry in Thailand. So you have this strange situation where the site is thoroughly confirmed by U.documents and journalistic reporting, but the host country still maintains a wall of silence.
Then there's the Salt Pit. This one's different from the others because it wasn't a purpose-built facility — it was a converted brick factory near Kabul, Afghanistan, that the CIA turned into a detention site. In November two thousand two, a detainee named Gul Rahman was stripped, shackled to a concrete floor, and left overnight in freezing temperatures. He died of hypothermia. The CIA's own inspector general investigated and referred the case for possible criminal prosecution, though the Justice Department declined to bring charges.
Rahman's family filed a lawsuit in U.federal court, and in twenty fifteen a judge allowed it to proceed — ruling that the case could go forward despite the government's state secrets claims. It was eventually dismissed on those grounds, but the fact that a federal judge even let it get to discovery tells you how strong the evidence was. The Salt Pit is not disputed. It's documented in the Senate report, in the CIA inspector general's files, and in court records.
That's five confirmed sites across four countries — Poland, Lithuania, Romania, Thailand, Afghanistan — each with a different kind of evidence chain, but all independently corroborated by at least two of the following: a court ruling, a parliamentary investigation, the Senate report, or declassified CIA cables.
Then there's the gray zone. For every confirmed site, there are a dozen alleged ones that remain stubbornly in the realm of the unverified.
This is where it gets tricky. The Témara detention facility, near Rabat, has been reported by Human Rights Watch, the New York Times, and multiple former detainees who describe being held there and interrogated by Morroccan intelligence with what they believed was U.But no court has ever confirmed it. No official investigation has substantiated it. The Morroccan government denies it existed. So you have credible journalism and consistent detainee testimony, but you don't have the kind of documentary or judicial evidence that exists for Stare Kiejkuty.
That's not to say Témara didn't exist — it probably did. But the evidentiary standard is fundamentally different, and that matters when you're trying to separate established fact from plausible-but-unproven claim.
Then there's the so-called Dark Prison near Kabul. This one is particularly haunting because it appears in the Senate report — multiple detainees described being held in a completely lightless facility where they were kept in solitary confinement, sometimes for months, with no sense of day or night. But investigators were never able to locate it. The detainees didn't know where they were — they'd been hooded during transport. The CIA has never acknowledged it. So you have this ghost facility, described consistently by multiple people, that simply cannot be pinned to a map.
Poland may have had a second site near Szymany airport, separate from Stare Kiejkuty. The Senate report hints at it, and some former officials have alluded to it, but it's never been confirmed by any investigation. So you see the pattern — the closer you get to the operational details, the fuzzier the picture becomes.
Which brings us to the strategic question. Why go to all this trouble? Why establish a global network of secret prisons, fly detainees across continents, involve foreign governments, create this enormous legal and diplomatic liability?
The CIA's logic, laid out in the two thousand two Office of Legal Counsel memos — the so-called Torture Memos — was essentially fourfold. First, legal black holes. If a detainee is held outside U.territory, the OLC argued, they have no constitutional habeas corpus rights. No right to a lawyer, no right to see a judge. The facility exists in a legal vacuum.
Second, operational security. By keeping locations unknown — even to other parts of the U.government — the CIA could control the flow of information. The Senate report documented multiple instances where the CIA misled the White House, Congress, and even the agency's own inspector general about what was happening at these sites.
Third, deniability for host governments. Poland and Lithuania could tell their publics and the European Union that they had no knowledge of what the CIA was doing on their soil — a claim that the ECHR rulings later demolished, by the way, but at the time it provided political cover.
Fourth, access to jurisdictions with weaker interrogation restrictions. If you're holding someone in a country where the local laws don't constrain what you can do, and the U.Constitution doesn't apply, you've created what the OLC memos essentially described as a permission slip for anything short of organ failure.
That's not a glib characterization. The memos explicitly defined torture as treatment that causes pain equivalent to "serious physical injury, such as organ failure, impairment of bodily function, or even death." If it didn't rise to that level, the OLC argued, it wasn't legally torture.
The strategic rationale was coherent, if you accept the premises. The question is whether it worked. And this is where the Senate report is absolutely devastating.
The report found — and this is based on the CIA's own internal documents, not outside criticism — that enhanced interrogation techniques produced unreliable information. Detainees would say anything to make the pain stop. The CIA's own Panetta Review, an internal assessment conducted in two thousand nine, concluded that traditional interrogation methods were more effective at producing actionable intelligence than EITs were.
The ticking time bomb scenario — the idea that you need to torture someone to get information that will prevent an imminent attack — the Senate report examined every case where the CIA claimed EITs had produced critical intelligence and found that in every single instance, the information had either been obtained through other means, was already known, or was simply false.
The CIA repeatedly misled policymakers about this. The report documents instances where the agency told the White House that a particular piece of intelligence came from EITs when internal cables showed it had actually come from a detainee who hadn't been subjected to them. It was a systematic misrepresentation of the program's effectiveness.
You have this enormous apparatus — secret prisons on three continents, a fleet of chartered aircraft, dozens of foreign government officials complicit — all built on a premise that the CIA's own internal reviewers later found to be false.
The legal arguments against black sites have been crystallized by the ECHR rulings, and they're worth understanding because they've fundamentally changed the landscape. The court established that states cannot contract out human rights violations. Poland and Lithuania argued that they weren't responsible because the CIA was running the facilities. The court said no — if you provide the building, knowing what it will be used for, you are complicit. You bear legal responsibility.
Convention Against Torture, Article Three, prohibits transferring anyone to a country where they face a real risk of torture. Extraordinary rendition — grabbing someone off the street and flying them to a black site in a country with no meaningful human rights protections — systematically violated that. is a signatory to the convention. The legal gymnastics required to argue that rendition didn't violate it were, to put it mildly, creative.
The ethical debate is harder to resolve cleanly. The proponents' argument isn't frivolous — if you genuinely believe that torture can prevent a mass-casualty attack, you're weighing the suffering of one detainee against the lives of potentially thousands of innocent people. That's a real moral dilemma, not a straw man.
The problem is that the empirical premise — that torture works — appears to be false. And even if it were true in some cases, you'd need a system that could reliably distinguish the cases where it works from the cases where it doesn't. The CIA's own experience suggests they couldn't make that distinction.
There's also the institutional corruption argument. Once you create a secret detention infrastructure, it develops its own momentum. The Senate report documented how the CIA continued using EITs long after the initial intelligence value had been exhausted, in part because the program had become self-justifying. Careers had been built on it. Budgets depended on it. Ending it would have meant admitting failure.
How should someone listening to this evaluate claims about black sites when they encounter them in the news or, more likely, in the darker corners of the internet where this topic tends to flourish?
I'd offer a three-question framework. First, has the site been confirmed by a court, an official investigation, or declassified documents? If the answer is yes — as with Stare Kiejkuty or the Salt Pit — you're on solid ground. If the answer is no, that doesn't mean the claim is false, but it does mean you're in a different evidentiary category.
Second, what was the facility actually used for? Detention, interrogation, transit, or something else? This matters because a lot of conspiracy content conflates everything into "secret torture dungeon," and that obscures real distinctions. A transit hub is different from an interrogation site, which is different from a long-term detention facility. Being precise about purpose helps you evaluate the claim.
Third, who's making the claim and what evidence do they provide? Multiple independent sources — journalistic, judicial, documentary — are much stronger than a single anonymous source. If the claim relies entirely on one person who says they heard something from someone who used to work somewhere, you should be skeptical. That doesn't mean dismiss it out of hand, but adjust your confidence accordingly.
The key legal legacy of all this is that the ECHR rulings have made it much harder for governments to host black sites today. The Polish and Lithuanian cases established that host nations bear legal responsibility, and that's not a theoretical risk — Poland actually paid compensation, and a Polish intelligence official was convicted. Any government considering a similar arrangement now has to weigh the diplomatic benefits against the very real possibility of criminal liability down the road.
For listeners who want to go deeper, the public record is surprisingly rich. The five-hundred-and-twenty-five-page executive summary of the Senate torture report is publicly available — it's the single best-documented source on the program. The ECHR rulings are public. New FOIA releases from the CIA's rendition program are still being declassified, and the Al-Nashiri case against Romania is still pending, which means there's more legal fact-finding to come.
Which leaves us with the big open question. Will the United States ever fully acknowledge all of the black sites? The CIA has never officially confirmed the existence of any specific facility — not Stare Kiejkuty, not Cat's Eye, not the Salt Pit — even though they've been litigated in European courts and documented in the Senate's own report. The state secrets privilege remains a powerful shield, and no administration, Democratic or Republican, has shown any appetite for lifting it on this issue.
The concept itself is evolving. As drone warfare and cyber operations replace physical detention as the primary tools of covert action, the classic black site may become obsolete. But the fundamental tension — between operational secrecy and legal accountability, between the desire to act without constraints and the rule of law — that's not going anywhere. The next generation of black sites might not be buildings at all.
They might be server farms, or algorithms, or jurisdictional arrangements we haven't even named yet. The form changes. The tension doesn't.
And now: Hilbert's daily fun fact.
Hilbert: In the high medieval period, around twelve hundred AD, Norse settlers in Greenland unknowingly built their settlements atop a subglacial lake. Modern ice-penetrating radar has mapped this buried lake at roughly twelve kilometers long, and if you were to compare its volume to something a medieval Greenlander would understand, it holds enough water to fill approximately forty million Viking-era longships.
...right.
That framework for evaluating black site claims is useful today, but the landscape is shifting. The question that lingers for me is whether the legal precedents set by the ECHR will actually constrain future behavior, or whether governments will simply find new forms of secrecy that the courts haven't caught up with yet. History suggests the latter, but the Polish conviction suggests the former isn't impossible.
That's the unresolved tension we're left with. The black sites of the early two thousands have been dragged into the light by courts and investigations, but the impulses that created them — the desire for legal black holes, for operational secrecy, for deniability — those haven't gone anywhere. Thanks to our producer Hilbert Flumingtop. This has been My Weird Prompts. If you want to dig into the primary sources we mentioned, the Senate report executive summary and the ECHR rulings are all publicly available — we'll link what we can on the website at my weird prompts dot com.
Until next time.
