MKUltra: What the CIA Actually Did, and What It Didn't

In the winter of 1953 a bacteriologist named Frank Olson checked into Room 1018A of the Statler Hotel in New York, in the company of a CIA doctor who was meant to be helping him. He had been anxious, sleepless, unlike himself, for more than a week. In the early hours of 28 November he went through the closed window and fell thirteen storeys to Seventh Avenue. The Agency called it a suicide brought on by overwork. What the family did not learn for another two decades was that nine days before his death, Olson had been served an after-dinner glass of Cointreau spiked with LSD by his own colleagues, as part of an experiment he had not agreed to and did not know was happening.

The programme that really existed

The temptation, when a story is this lurid, is to assume it has been embroidered. It has not been embroidered enough. MKUltra was a real, authorised, funded programme of the United States Central Intelligence Agency, and almost everything solid we know about it we know because the government’s own paperwork survived by accident.

It began formally in April 1953, approved by Director Allen Dulles, run out of the Agency’s Technical Services Staff. The man in charge was a chemist named Sidney Gottlieb — a rural smallholder and folk-dance enthusiast with a club foot and a stammer, who kept goats and made his own yoghurt, and who authorised some of the most disturbing human experimentation in American peacetime history. The premise was Cold War fear made operational: the belief that the Soviets, the Chinese and the North Koreans had learnt to break and rebuild the human mind, and that the United States could not afford to be behind.

MKUltra was an umbrella. Beneath it ran at least 149 subprojects, farmed out to universities, hospitals, prisons and private researchers, most of whom had no idea whose money they were spending because it was laundered through a front called the Society for the Investigation of Human Ecology. Subproject 68 sent money to the psychiatrist Ewen Cameron at the Allan Memorial Institute in Montreal, where patients who had come in for anxiety or post-natal depression were subjected to “psychic driving” — drugged into weeks of induced sleep, dosed with LSD, and made to listen to looped recorded messages hundreds of thousands of times while their existing personalities were, in Cameron’s own word, “depatterned”. Some never recovered. Decades later the Canadian government quietly settled with their families.

The tool the programme kept returning to was LSD. Gottlieb was convinced it might be a truth serum, an interrogation key, a way to make an enemy agent malleable — and to understand it, his staff dosed each other, dosed colleagues, and dosed strangers who had no notion they had been drugged. Frank Olson was one of the colleagues. The strangers came later.

Midnight Climax

The most cinematic and the best-documented corner of the programme was run by a former Federal Bureau of Narcotics agent named George Hunter White, under the sub-heading Operation Midnight Climax. From the mid-1950s the CIA maintained safehouses — one in Greenwich Village, others in and around San Francisco and Marin County — fitted out as ordinary apartments and wired for observation. White paid prostitutes to bring men back to these flats, where the men were dosed with LSD without their knowledge while White sat behind a two-way mirror with a jug of martinis, watching what the drug did to a person who had no idea he had taken it.

White was not a tortured man. In a letter to Gottlieb near the end of his life he wrote, with a candour that has become one of the most quoted lines of the whole affair, that it had been “fun, fun, fun” — where else could a red-blooded American boy lie, kill and cheat with the sanction of the all-highest? That sentence tells you something the redacted files do not: that the mundane machinery of a democratic government had produced a space where a man could do these things to strangers and feel it as a lark.

We know all of this because of a paperwork accident. In 1973, with the Watergate scandal metastasising and congressional investigators beginning to circle, Director Richard Helms ordered the MKUltra files destroyed. Gottlieb complied; the operational records were shredded. Helms and Gottlieb both later testified that they believed the entire documentary trail had been eliminated. It had not. A cache of roughly 20,000 pages of financial records — expense claims, invoices, the dull accountancy of secret work — had been misfiled in a separate storage system and escaped the shredder. In 1977 those pages were recovered under a Freedom of Information Act request filed by the writer John Marks, and the money trail was what let the Senate reconstruct what had been done — the invoices told the story the shredded operational files were meant to bury.

What the record actually shows

Here is the part that matters most, because it is the hinge on which the myth turns. Read the surviving files and the 1977 Senate testimony carefully, and a very specific conclusion emerges: MKUltra was real, it was criminal, it ruined and ended lives — and it did not work.

The Agency spent the better part of a decade and a great deal of money trying to find a chemical route to controlling a human mind, and it failed. LSD did not produce reliable, directable behaviour; it produced chaos, terror and unpredictability. Nobody was turned into a programmable assassin. Nobody was made to betray a secret they would otherwise have kept in any way that could be trusted. The Church Committee and the 1977 Senate Select Committee hearings, chaired by Ted Kennedy, laid out a programme that was a moral catastrophe and an operational dead end at the same time. The horror of MKUltra lies in its futility: people were destroyed in the pursuit of a power that was never achievable.

It is worth being concrete about why the science failed, because the failure is central to everything that follows. LSD is a disinhibitor and a disorganiser of experience; it dissolves the ordinary boundaries of perception and self. What an interrogator needs is the opposite — reliability, direction, a subject who can be steered toward a predetermined behaviour and depended upon to remember or forget on command. A drug that makes a person suggestible in unpredictable directions is useless for control, because you cannot govern where the suggestibility leads. Gottlieb’s team learnt this the expensive way, dosing subject after subject and watching the results scatter. By the early 1960s the Agency’s own enthusiasm had cooled; MKUltra was folded into a successor programme, MKSearch, and quietly wound down. The dream of the chemical key had died inside the CIA years before the public ever heard its name.

This is the well-sourced, dated, documented kernel: the human experimentation of a paranoid decade, recovered from financial ledgers because the operational files were deliberately burned.

Where the record ends and the story takes over

Now watch the fork. The proven programme has a boundary, and the boundary is precisely the point at which the mythology begins.

The Agency tried to control minds and could not. The mythology holds that it can — that somewhere behind the failed public programme a successful secret one exists, capable of manufacturing killers to order. The film The Manchurian Candidate, released in 1962 while MKUltra was still running, supplied the template a full fifteen years before the public knew the real programme existed: the ordinary man with a hidden trigger, the assassin who does not know he is one. When the Senate hearings broke in the late 1970s, the fiction was already installed in the culture, waiting to be baptised as documentary.

From there the claim metastasised in a predictable direction. Nearly every American mass shooting of the past forty years has, within days, attracted the assertion that the perpetrator was an MKUltra product, a “sleeper” activated by a handler. Sirhan Sirhan’s defence has invoked it; the lone gunmen of a dozen atrocities have been retrofitted with imaginary CIA controllers. The reasoning runs backwards from an unbearable event: a young man walks into a school and does something no human story seems able to explain, and the mind reaches for an author. If he was programmed, then the horror has a designer, and a designer means the chaos was really order all along.

Why the myth is stronger than the truth

The mythology is durable for a reason that has nothing to do with evidence, and it is worth naming plainly, because it is the same reason MKUltra as history is so hard to believe when you first meet it.

The people who suspected, in 1955, that a branch of their own government was dosing citizens with hallucinogens in wired apartments would have been called paranoid. They were right. That is the corrosive gift MKUltra left behind: proof that the paranoid position was, on one specific and documented occasion, the accurate one. Once a government has been caught doing the unthinkable, “they would never do that” stops being a usable argument, and every subsequent claim, however unmoored, gets to borrow the credibility of the one that turned out to be true.

The believer in the programmed shooter is not being stupid. They are performing a reasonable-looking inference from a real premise — the state secretly experimented on minds — to an unreal conclusion — therefore it can manufacture behaviour at will. The move from tried and failed to succeeded in secret is the entire error, and it is almost invisible, because the first half is so much worse than most people are prepared to believe that the second half arrives feeling like more of the same.

There is a cousin to this pattern worth sitting with. The most powerful conspiracy theories are rarely pure invention; they are usually a documented outrage with an extra storey built on top. The Tuskegee syphilis study earned Black America’s medical mistrust with forty years of real, sanctioned neglect, and that earned mistrust then attached itself to claims the record does not support. The FBI’s COINTELPRO operations proved that activists who said they were being watched were telling the truth, which made every later surveillance fear feel confirmed in advance. MKUltra belongs in that family. The kernel is genuine and it is damning, and the mythology grows precisely because the kernel is too dark to dismiss.

The honest thing to hold, at the end, is both halves at once and in the right proportion. The CIA drugged unwitting people, drove some of them into breakdown, contributed to at least one death, and burned the evidence — and it never learnt to control a single mind. Frank Olson’s family fought for decades to have the fall in New York recognised for what the paperwork eventually showed it to be. The tragedy of MKUltra is complete without the Manchurian candidate. It never needed him. He is what we added, because a designed horror is somehow easier to carry than the true one, which is that a serious government spent years and lives chasing a power that does not exist, and hurt real people getting nowhere.