Joseph Weizenbaum

Overview

Joseph Weizenbaum (1923–2008) was a German-born American computer scientist whose career took an unusual shape: he built a program that could carry on a written exchange with a person, and then spent much of his later working life arguing about what such a program does and does not show ^[2][3]. He was born in Berlin on 8 January 1923 ^[1] and died in Berlin on 5 March 2008 ^[2]. He was a fellow of the American Association for the Advancement of Science ^[2].

Education and early career

Weizenbaum left Germany as a child, fleeing the Nazi regime with his parents and arriving in the United States in the mid-1930s ^[2]. He earned a BS in 1948 and an MS in 1950, both from Wayne University ^[1].

From 1955 to 1963 he was a systems engineer at the Computer Development Laboratory of General Electric ^[1]. There he was part of the team that designed and built the first computer system dedicated to banking operations ^[2]. This was a period in which a large computer was usually planned and built around the work of one organisation rather than bought off the shelf, so a bank's operations were an early test case for that kind of dedicated design.

ELIZA

In 1966 he published ELIZA in Communications of the ACM, and the program operated through a script called DOCTOR ^[3]. The design separated the program from the text that governed its replies: the conversation was shaped by the script rather than written into the program itself, and the paper presented the program together with that script ^[3].

What made the program interesting to readers was not that it reasoned about anything, but that a person typing at a terminal could receive replies in return and read a conversation into them. That gap between what the machine does and what the user experiences became the subject Weizenbaum returned to in his later writing.

A critic of artificial intelligence

In 1976 he published the book Computer Power and Human Reason: From Judgment to Calculation, an influential critique of artificial intelligence ^[2][3]. The subtitle sets judgment against calculation, and the book's argument turns on where the line between the two should be drawn: which decisions, in medicine, in government or in ordinary life, call for human judgment and should not be handed over to a computer even when a computer could produce an answer. The book gave him a public voice in the debate over what computers should be asked to do, and it is the work for which he is remembered outside the technical literature as much as ELIZA is inside it ^[2][3].

At MIT

He joined MIT in 1963 as a visiting associate professor of computer science ^[2], and in 1970 he became professor of computer science and engineering ^[1]. MIT was his base for both the ELIZA work and the book that followed it ^[2][3].

Influence

Weizenbaum's reputation rests on a pair of contributions that pull in opposite directions. ELIZA opened a line of work on programs that hold conversations with people, and its script-driven design made the point that a convincing exchange does not require the machine to understand anything ^[3]. Computer Power and Human Reason made the opposite point, that the readiness to hand human questions to calculation is a choice with consequences, and it is still cited as an influential statement of that case ^[2][3]. Taken together they describe a career spent close to the technology and uneasy about the conclusions drawn from it.