PARRY (short for
paranoia) is the first computer program to imitate a mental patient, created by
Stanford University psychiatrist Kenneth Colby.
PARRY is communicated with by the psychiatrist-user in
simple English.
PARRY's responses are intended to mirror the cognitive
(mal)functioning of a paranoid patient.
In the late 1960s and early 1970s, Colby experimented with
mental patient chatbots, which led to the development of PARRY.
Colby sought to illustrate that cognition is fundamentally a
symbol manipulation process and that computer simulations may help psychiatric
research.
Many technical aspects of PARRY were shared with Joseph
Weizenbaum's ELIZA.
Both of these applications were conversational in nature,
allowing the user to submit remarks in plain English.
PARRY's underlying algorithms, like ELIZA's, examined
inputted phrases for essential terms to create plausible answers.
Parry, who was fictitious, was a gambler who had gotten into
a fight with a bookie.
Parry was paranoid enough to assume that the bookie would
send the Mafia after him.
Since a result, PARRY freely shared information on its crazy
Mafia ideas, as it would wish to enlist the user's assistance.
PARRY was also born with the ability to be "sensitive
to his parents, religion, and sex" (Colby 1975, 36).
In most other topics of conversation, the show was neutral.
If PARRY couldn't find a match in its database, it may
respond with "I don't know," "Why do you ask that?" or by
returning to an earlier subject (Colby 1975, 77).
Whereas ELIZA's achievements made Weizenbaum a skeptic of
AI, PARRY's findings bolstered Colby's support for computer simulations in
psychiatry.
Colby picked paranoia as the mental state to mimic because
it has the least fluid behavior and hence is the simplest to see.
Colby felt that human cognition was a process of symbol
manipulation, as did artificial intelligence pioneers Herbert Simon and Allen
Newell.
PARRY's cognitive functioning resembled that of a paranoid
human being as a result of this.
Colby emphasized that a psychiatrist conversing with PARRY
had learnt something about human paranoia.
He saw PARRY as a tool to help novice psychiatrists get
started in their careers.
PARRY's reactions might also be used to determine the most
successful therapeutic discourse lines.
Colby hoped that systems like PARRY would assist confirm or
refute psychiatric hypotheses while also bolstering the field's scientific
credibility.
On PARRY, Colby put his shame humiliation hypothesis of
paranoid insanity to the test.
In the 1970s, Colby performed a series of studies to see how
effectively PARRY could simulate true paranoia.
Two of these examinations resembled the Turing Test.
To begin, practicing psychiatrists were instructed to
interview patients using a teletype terminal, an antiquated electromechanical
typewriter that was used to send and receive typed messages over
telecommunications.
The doctors were unaware that PARRY was one of the patients
who took part in the interviews.
The transcripts of these interviews were then distributed to
a group of 100 psychiatrists.
These psychiatrists were tasked with determining which
version was created by a computer.
Twenty psychiatrists properly recognized PARRY, whereas the
other twenty did not.
A total of 100 computer scientists received transcripts.
32 of the 67 computer scientists were accurate, while 35
were incorrect.
According to Colby, the findings "are akin to tossing a
coin" statistically, and PARRY was not exposed (Colby 1975, 92).
Find Jai on Twitter | LinkedIn | Instagram
You may also want to read more about Artificial Intelligence here.
See also:
Chatbots and Loebner Prize; ELIZA; Expert Systems; Natural Language Processing and Speech Understanding; Turing Test.
References & Further Reading:
Cerf, Vincent. 1973. “Parry Encounters the Doctor: Conversation between a Simulated Paranoid and a Simulated Psychiatrist.” Datamation 19, no. 7 (July): 62–65.
Colby, Kenneth M. 1975. Artificial Paranoia: A Computer Simulation of Paranoid Processes. New York: Pergamon Press.
Colby, Kenneth M., James B. Watt, and John P. Gilbert. 1966. “A Computer Method of Psychotherapy: Preliminary Communication.” Journal of Nervous and Mental Disease 142, no. 2 (February): 148–52.
McCorduck, Pamela. 1979. Machines Who Think: A Personal Inquiry into the History and Prospects of Artificial Intelligence, 251–56, 308–28. San Francisco: W. H. Freeman and Company.
Warren, Jim. 1976. Artificial Paranoia: An NIMH Program Report. Rockville, MD: US. Department of Health, Education, and Welfare, Public Health Service, Alcohol, Drug Abuse, and Mental Health Administration, National Institute of Mental Health, Division of Scientific and Public Information, Mental Health Studies and Reports Branch.
