Last Updated on December 8, 2020 by Sean B
The history of machine intelligence is not merely limited to people whose only job is to write codes in machine language algorithms; instead, there’s a diverse professional sphere within the field of artificial intelligence that has its philosophical and psychiatric implications. Specifically, the notion of consciousness is often discussed in debates about Artificial Intelligence, and many people have defied our expectations about what an A.I. program can achieve.
In this blog, we look at someone who helped discuss machine intelligence forward, Kenneth Colby. We will provide a brief biography of Kenneth Colby dealing with his studies, career, and contributions to Computer Science and Artificial Intelligence.
Early Life and Studies
Kenneth Colby was born in Waterbury, Connecticut, the USA, in 1920. He graduated from Yale University in 1941 and received his M.D. from Yale Medical School in 1943.
Career and Research
Kenneth Colby started practicing psychoanalysis for the first several decades of his career. He was a clinical associate at the San Francisco Psychoanalytic Institute in 1951. During his work as a psychoanalyst, he published a primer called “An introduction to psychodynamic psychotherapy.”
Colby later joined the Department of Computer Science at Stanford University in the 1960s. During his time at the Department of Computer Science, Kenneth started his pioneering work in Artificial Intelligence, which was a relatively new field at that time.
His pioneering was recognized on an academic level by the National Institute of Mental Health when he was awarded a Career Research Scientist Award in 1967. Colby then joined UCLA as a psychiatry professor in 1974 and then soon became a jointly appointed professor in the Department of Computer Science.
Dr. Colby wrote numerous works such as books and articles on psychiatry, psychology, psychotherapy, and Artificial intelligence during his lifetime as a professor and researcher.
Kenneth Colby As a Psychoanalyst
As a psychoanalyst, Kenneth Colby’s work consisted of reconciling Sigmund Freud’s ideas with modern physics concepts and science’s philosophy. Kenneth published his book, “Energy and Structure in Psychoanalysis,” in 1951 based on that goal.
However, Kenneth soon forfeited his attempts to reconcile Freud’s work with modern science after seeing contemporary science and psychoanalysis developments. Kenneth became disillusioned with Freud’s work and psychoanalysis in general when he considered the pseudo-scientific aspects of it.
Kenneth expresses his disillusion with psychoanalysis later in his publications, such as his 1958 book, a skeptical psychoanalyst. His rigorous psychoanalysis criticism remained apparent throughout his works, such as his 1983 book, Fundamental Crisis in Psychiatry.
Kenneth’s later views and research in psychiatry was heavily based on his experience with psychoanalysis.
Kenneth Colby As a Computer Scientist
Kenneth Colby, in 1960 began thinking about the possible ways in which the Computer Theory and application will be able to contribute towards the understanding of human brain functions and mental illnesses.
Kenneth worked on his initial project that consisted of developing an intelligent speech prosthetic that would help patients suffering from aphasia speak by searching and articulating words on their behalf from phonemic or semantic they can generate.
Kenneth Colby also was among the first to consider the endless possibilities of computer-assisted psychotherapy. With his son Peter Colby, Kenneth Colby formed the company Malibu Artificial Intelligence Works to develop and market a natural language version of cognitive-behavioral therapy for depression in 1989.
They called that natural language “Overcoming Depression.” Overcoming depression was then used by the U.S. Navy and the Department of Veteran Affairs as a therapeutic learning program. It then was distributed to individuals who could use it without the supervision of a psychiatrist.
However, this program was looked down on in the media and was denigrated. Colby then said to a journalist that this program will have better professional ethics standards and will not get tired or burn out, unlike a human.
Kenneth Colby in Artificial Intelligence
In the 1960s, When Colby was with Stanford University, he became fascinated with chatbots, a relatively new idea. Kenneth Colby started working on and developing chatterbots.
Kenneth Colby decided to develop a chatbot that could hold a conversation with a person. A very well-known chatbot was ELIZA, a chatbot created by Joseph Weizenbaum at MIT’s artificial intelligence laboratory in 1966.
ELIZA was moduled after a Rogerian psychotherapist and was meant to simulate the experience between a human agent and a normal psychotherapist. ELIZA was the first chatterbot of its time, and due to its relevance to psychiatry, it influenced Kenneth’s work.
Joseph Weizenbaum designed his chatbot after George Bernard Shaw’s character “ELIZA Dollitle” from “Pygmalion.” Joseph Weizenbaum had never expected his program to gain so much recognition as he thought that machines would never be able to replace humans.
Kenneth Colby then built upon ELIZA’s idea to create a similar Stanford Artificial Intelligence Laboratory program in 1972. Colby designed a natural language program called PARRY that could stimulate the mental state of suffering from schizophrenia.
PARRY stimulates the patient through a complex system of assumptions, attributions, and emotional responses triggered by shifting weights assigned to verbal inputs.
PARRY – The Paranoid Chatterbot
The most significant contribution of Kenneth Colby to Artificial Intelligence was the development of PARRY, a chatterbot stimulating a patient who has schizophrenia. The thinking of PARRY involved him misrepresenting the motives of others.
He would doubt other people’s motives and intentions and think that other people are concealing some ill-intention. One of Kenneth’s motivations for developing PARRY was to create a virtual teaching system for medical students in which they could interact before performing on a human patient.
PARRY was the embodiment of what Kenneth Colby saw defines paranoia. For Kenneth Colby, paranoia was a form of degeneration that emerged due to not processing and interpreting symbols accurately.
For Kenneth Colby, this degeneration could be reprogrammed or cured since the underlying problem was tied to the algorithm inside the brain or the processor. After PARRY was introduced, it immediately sparked a lot of controversy and debate on the nature and implications of Machine intelligence and its role in psychiatry.
Kenneth Colby claims that PARRY stimulates a paranoia patient by mimicking an actual paranoia patient’s human responses. Colby said that the program’s internal structure is similar to the brain structure of an average human.
Colby said that although we are currently unaware of how a normal human brain functions, the structure that allows PARRY to make linguistic decisions is the theoretical proximity of the human brain structure.
PARRY is considered the first program to have won the Turing test. Turing test was a test first proposed by a pioneer in cryptoanalysis, computer science, and mathematics, Alan Turing. Turing test involves a chatbot holding a conversation with a judge.
If the chatbot can successfully fool the judge into thinking that he is talking to a human and not a machine, the bot will pass the Turing test. When PARRY took the Turing test, judges could not distinguish PARRY from human agents suffering from paranoia.
PARRY participated in the variation of a Turing test in the 1970s. A group of professional psychiatrists was invited to analyze a team of real patients suffering from paranoia and an array of computers running PARRY through teleprinters.
A group of 33 psychiatrists was shown transcripts of the conversation between the judges and PARRY, the chatbot. The two groups were asked to identify which one was a human and which one was the chatbot.
It was found that the judges were only able to guess the correct answer 48% of the time, which is merely equivalent to random guessing. The American cognitive scientist and philosopher Daniel Dannet called Kenneth Colby the only serious participant to attempt to win the Turing test.
A notable difference between ELIZA and PARRY is that ELIZA was criticized for its lack of an internal world model that could track the conversations’ flow. However, PARRY was able to generate his paranoid responses by following his internal emotional, mental states.
ELIZA Meets PARRY
Joseph Weizenbaum’s chatbot ELIZA, the psychiatrist, has often engaged in a PARRY conversation; the chatbot is stimulating a paranoia patient. The most notable example of their meeting occurred at the ICCC 1972. During the event, ELIZA and PARRY were hooked up over ARPANET and engaged in a conversation.
The Significance of PARRY
Even though PARRY has sparked a lot of controversy and debate, his significance and achievements in the domain of artificial intelligence remain apparent. As in 1999, Yorick Willis and Roberta Catizone from the University of Sheffield described a conversation with PARRY in their paper Human-Computer Conversation:
“The best performance overall in HMC (Human-machine conversation) has almost certainly been Colby’s PARRY program since its release on the (then ARPA) net around 1973. It was robust, never broke down, always had something to say and, because it was intended to model paranoid behaviour, its zanier misunderstandings could always be taken as further evidence of mental disturbance, rather than the processing failures they were.”
Other Works and Areas of Interests
Besides computer science and artificial intelligence, Kenneth Colby also developed an interest in other anthropology areas, where he explored the sex differences in the dreams of primitive tribes.
Along with that, Kenneth Colby was interested in chess and wrote a book, “the secrets of a Grand patzer” on chess.
Kenneth Colby died on the 20th of April in 2001.
Dr. Kenneth Colby made significant contributions to the fields of both Artificial Intelligence and Psychiatry during his life. The PARRY chatbot is one of the most successful chatbots ever created, and remains incredibly influential today. Colby managed to give PARRY a personality by including an entirely human flaw, mental illness. This inclusion of flaws (and other human features) continues to be an important feature in chatbots, influcencing the creation of chatbots like Eugene Goostman and HeX among others.
Colby’s chatbot PARRY was a huge leap forward and the meeting between ELIZA and PARRY remains one of the more interesting events in chatbot history.
We hope that the blog was insightful.
We are looking forward to reading your thoughts in the comments.