The CIA secretly experimented on people with drugs and mind control

What they actually did

Across its two decades (from Dulles's 1953 authorisation to the programme's formal end in 1973), the obsession, above all, was LSD. Under chemist Sidney Gottlieb, who ran MKUltra's day-to-day operations for the CIA, researchers dosed themselves, dosed each other, and (the detail that turns a research scandal into a civil-liberties one) dosed unwitting members of the public, then studied what happened.

In Operation Midnight Climax, the agency ran safehouses at 225 Chestnut Street in San Francisco and in New York City through the mid-1950s and into the 1960s, where sex workers on the CIA's payroll lured men back to the apartments, slipped them LSD, and let agents watch through two-way mirrors. The operation was overseen by a former federal narcotics agent, George Hunter White, working under the cover name “Morgan Hall” in close coordination with Gottlieb, until a 1963 report by the CIA's own inspector general recommended shutting the safehouses down.

The darkest documented chapter is the death of Frank Olson, an Army biochemist. At a CIA retreat at Deep Creek Lake, Maryland, on the evening of 19 November 1953, Gottlieb slipped LSD into his drink without his knowledge. Nine days later, in the early hours of 28 November, Olson fell from the tenth-floor window of Manhattan's Hotel Statler and died on the pavement below. His family was told he had suffered a breakdown; only in 1975, after the Rockefeller Commission's report forced the CIA to admit the covert dosing, did they learn the truth. President Ford invited them to the White House to apologise in person, calling what had been done to Olson “illegal and unconscionable,” and in 1976 Congress voted the family a $750,000 settlement. A 1994 exhumation the family commissioned, led by forensic scientist James Starrs, found a skull injury the examining team concluded had occurred before Olson went through the window, not during it; evidence Starrs called “rankly and starkly suggestive of homicide.” Officially, the government's account has shifted from suicide to “death by misadventure,” but never to homicide.

That any of this can be stated with such precision is almost an accident of filing. As the programme wound down in 1973, Gottlieb and Director Richard Helms agreed the MKUltra records should be destroyed, and the retiring Gottlieb personally drove to the CIA's own records centre and ordered its staff to burn the files. What survives today survives only because a separate cache of budget paperwork had been misfiled elsewhere and was overlooked that same year, then forced into public record by the Church Committee's 1975 findings and a 1977 Senate hearing. The paper trail behind nearly every fact above exists because of that one clerical accident.

The programme was not testing a theory about strangers. It was testing drugs on citizens who never agreed to take part, including, in at least one documented case, a colleague who paid for it with his life.

Why a proven conspiracy matters

MKUltra is the reason a comforting sentence (“the government would never do that”) stopped being safe to say. That is its real legacy, and it is a legitimate one.

It is also, for exactly that reason, the most-cited exhibit in the case for theories that have no evidence at all. If they really dosed civilians with LSD, the argument runs, then why not this, why not that? The leap is understandable and usually wrong: a proven abuse in one place is not evidence for an imagined one somewhere else.

The honest lesson of MKUltra is narrower and more useful. Institutions really do sometimes conspire against the people they answer to, which is precisely why every claim deserves its own evidence, its own paper trail and its own verdict, rather than a borrowed one.

For twenty years the answer to “would they really do that?” was no. MKUltra is the file that turned it into: sometimes.