PROJECT BLUEBIRD

Opening

Project BLUEBIRD was the CIA's first behavioural control programme. On 5 April 1950 the Agency's Security chief, Sheffield Edwards, proposed interrogation teams built around "the polygraph, drugs, and hypnotism". Director Roscoe Hillenkoetter approved it on 20 April 1950 and authorised unvouchered funds for its most sensitive parts. The stated aim was defence: learn how communist states seemed to break prisoners, and protect American personnel from the same methods.

The record shows more than defence. Within three months a team was in Tokyo dosing subjects with sodium amytal and benzedrine and trying to induce amnesia. By October 1950 it had used "advanced" techniques on 25 people, apparently North Korean prisoners of war. In August 1951 BLUEBIRD was renamed ARTICHOKE, the line of work that led to MKUltra in April 1953.

Timeline

February 1949Hungary tries Cardinal József Mindszenty, who confesses to treason with what Marks calls a glazed look in his eyes. A CIA security memorandum says "some unknown force" controlled him. The fear that Moscow had a method drives everything after.

Summer 1949The Agency's head of Scientific Intelligence travels to Western Europe "to apply special methods of interrogation for the purpose of evaluation of Russian practices", and recommends a team trained in such methods.

5 April 1950Edwards's office memorandum to Hillenkoetter sets out the project, its training plan and a budget of $65,515. It asks that knowledge of BLUEBIRD be "restricted to the absolute minimum number of persons."

20 April 1950Hillenkoetter approves BLUEBIRD.

9 May 1950A planning meeting asks for the Nuremberg Trials papers to be searched for material on drugs, narcoanalysis and special interrogation.

July 1950The first three-man team reaches Tokyo, about a month after the Korean War begins. Four subjects, probably suspected double agents, receive sodium amytal with benzedrine; the last two also get picrotoxin.

About October 1950The team uses "advanced" techniques on 25 subjects, apparently North Korean prisoners.

End of 1950Morse Allen of the Office of Security becomes head of BLUEBIRD.

2 February 1951A memo asks about obtaining six "hypospray" jet injectors and adapting a tear gas pencil.

20 August 1951The project is renamed ARTICHOKE.

April 1953MKUltra is approved and takes over much of the research.

1973Many related records are destroyed on Director Richard Helms's order.

1977Surviving BLUEBIRD and ARTICHOKE papers come out through FOIA requests and the Senate hearings that followed the discovery of MKUltra financial records.

Key figures

Sheffield Edwards, the former Army colonel who ran the CIA's Office of Security and proposed the teams.

Roscoe Hillenkoetter, Director of Central Intelligence, who approved the programme and its unvouchered funding.

Morse Allen, the Security officer who headed BLUEBIRD from late 1950 and pushed hard on hypnosis.

The subjects in Japan, never named in any released document.

John Marks, whose 1979 book The Search for the "Manchurian Candidate" first reconstructed the programme from CIA papers.

Physical and documentary trail

The founding memorandum of 5 April 1950 survives and now sits in the National Security Archive's collection "CIA and the Behavioral Sciences", released between December 2024 and April 2025 with more than 1,200 records. It specifies three-man teams: a doctor, ideally a psychiatrist, a polygraph operator trained in hypnosis, and a technician. According to Marks, Edwards agreed not to use them operationally without the permission of a high-level committee.

The Church Committee's Book I (1976) summarises the programme's objectives, records that they widened to include offensive uses, and dates the renaming to ARTICHOKE in August 1951. The Office of Scientific Intelligence led at first; responsibility passed to the Inspection and Security Office in 1952. An August 1951 meeting record in the 2025 release refers to a "project in Japan and Korea" in which the Army used a polygraph operator with psychiatrists and psychologists on Korean prisoners.

Marks summarised the goal that kept recurring in the papers of this period:

controlling an individual to the point where he will do our bidding against his will and even against such fundamental laws of nature as self-preservation.

Corrections to the earlier version of this report. No factual errors were found. The earlier version left out the 5 April proposal, Edwards's role and the budget, and gave the Japan tests no dates; those are added here from Marks and the 2025 release. Many operational details were lost in the 1973 destruction, and the release fills some gaps, not all.

Things you might not know

• The cover story was a lie detector. The Tokyo team was told to hide its purpose even from the US military and to describe its work as "intensive polygraph" sessions.
• The budget was $65,515. It covered salaries and kit down to syringes, towels and film cameras, with $18,000 for transport to destinations still redacted.
• One officer called the bluff in writing. On reading one of the memos he noted that if it was supposed to be covered up as a defensive feasibility study, it was "pretty damn transparent".
• They shopped for gadgets. The February 1951 memo also asked about a German "Scheintot" pistol, a device meant to make a target look dead.
• The name meant nothing. Marks notes that BLUEBIRD, like other Agency code names, had no significance except perhaps to the person who chose it.
• Nuremberg was a dead end. A CIA psychologist who read the German experiments told Marks they were "a real horror story" but of little specific help.

Competing explanations

A defensive programme. The CIA's stated reason was to understand and counter Soviet methods after Mindszenty and the other show trials. The founding memo frames it as protection of Agency personnel and the checking of agents and defectors, and that is part of the record.

An offensive programme from the start. The Japan tests were run on prisoners and suspected agents who could not refuse, to develop techniques the CIA could use itself. Marks concluded that the line between offence and defence soon became so blurred as to be meaningless.

A response to a threat that was not there. A Gottlieb deposition from 1983, cited in 2026 reporting on the release, recalls a finding that Chinese and Korean methods with American prisoners were not drug-based or esoteric. On this reading BLUEBIRD chased a capability its enemies did not have.

Later life of the case

BLUEBIRD is usually told as the first chapter of Project MKUltra, and for decades Marks was the only public source on the prisoners in Japan. In April 2026 The Intercept reported on the National Security Archive release, which confirms that Korean prisoners in American custody were tested and that further work was planned. It established a pattern that ran through ARTICHOKE and MKUltra: testing on people who were in no position to say no.

Finding

BLUEBIRD is admitted and documented. The CIA approved it in April 1950, funded it off the books, and within months tested drugs and amnesia techniques on prisoners and suspected agents in Japan. Defence and offence both appear in the papers, and the offensive aim is stated plainly.

What still will not resolve: who the subjects were. Four people in July 1950 and 25 more that autumn are counted in the record, but none is named, and nothing released says what happened to them afterwards.