Kenneth Mark Colby was a psychiatrist who became one of the early figures in conversational AI. His 1971 paper “Artificial Paranoia,” published in the journal Artificial Intelligence with co-authors Sylvia Weber and Franklin Dennis Hilf, identifies him as a Senior Research Associate in the Computer Science Department at Stanford University. The work was supported in part by the National Institute of Mental Health and the Advanced Research Projects Agency, reflecting both its clinical motivation and its place in the early AI research community.
Colby’s signature creation was PARRY, a program that modeled the conversational behavior of a paranoid patient. Coming from psychiatry rather than engineering, he approached the chatbot not as entertainment but as a tool for studying belief, affect, and the mechanics of paranoid thinking, and as something a trainee might practice on before meeting real patients. He proposed evaluating the model with indistinguishability tests, in which expert judges tried to tell the program’s transcripts from those of actual patients.
His perspective stood in contrast to that of Joseph Weizenbaum, whose ELIZA simulated a therapist and who grew deeply skeptical of treating such programs as if they understood people. Colby, by contrast, defended the scientific value of building and testing computational models of disordered cognition.
For a general reader, Colby illustrates how the history of conversational systems is entangled with medicine and the science of the mind, and how questions about whether software can convincingly imitate a human were being posed by clinicians, not only computer scientists.