This Day in Jewish History

1923: A Scientist Who Would Quail at the Artificial Intelligence He Created Is Born

Joseph Weizenbaum created Eliza, software 'human' enough to frighten him badly: Computers aren't wise, he reasoned, and shouldn't be given tasks requiring wisdom.

Prof. Hans Uszkoreit (left) and Prof. Joseph Weizenbaum, creator of the "smart" software Eliza, which convinced some people it was human - which led the professor to mull the potential danger of artificial intelligence.
Andreas Schepers, Flickr

January 8, 1923, is the birthdate of Joseph Weizenbaum, a German-born computer science pioneer, whose early work in the field of artificial intelligence led him to become apprehensive about its negative potential. After he created Eliza, a computer program that could engage in seemingly intelligent conversation with a human, Weizenbaum became something of a Cassandra, warning the public about the dangers he thought AI constituted for society.

Joseph Weizenbaum was born in Berlin, in Weimar Germany, to Yechiel and Henrietta Weizenbaum. Jechiel was a furrier, who, Joseph revealed in a documentary film about him made in 2007, caused him to have feelings of inferiority, particularly vis-à-vis his charismatic older brother, Heinz: "My father was absolutely convinced that I was a worthless moron, a complete fool, that I would never become anything," he recounted in “Joseph Weizenbaum: Rebel at Work.” 

In 1936, as the Nazi noose became tighter for German Jews, the family emigrated to the United States, settling in Detroit, Michigan.

Those early years in Germany had a profound effect on Weizenbaum’s outlook: Along with suffering from depression during his adult life, he was especially alert to the threat of future totalitarian regimes like the Nazis, and the use they could make of intelligence technology.

The unwitting therapist

In 1941, Joseph began studying at Wayne University (today Wayne State), in Detroit, interrupting his studies the following year to join the U.S. Army Air Forces, serving as a meteorologist.

(Joseph’s brother Heinz, in the meantime, converted to Roman Catholicism, changed his name to Henry Sherwood, and, after army service in the Pacific, became an early computer scientist who worked with a CIA contractor behind the Iron Curtain.)

Joseph Weizenbaum, 5th from the right, at the World Economic Forum in Davos, Switzerland, in 2008.
Gisiger, Flicker Creative Commons

Joseph received his B.S. in math from Wayne University in 1948, and two years later a master’s. After working briefly as a research assistant at the university, helping to build a digital computer, Weizenbaum began working for General Electric in 1956, which at the time had a contract to develop an early computer system for the Bank of America. Among other things, he invented the fonts that appear at the bottom of bank checks, that are read by Magnetic Ink Character Recognition.

It was in 1966, while working as a visiting professor at MIT, that Weizenbaum published the computer program “Eliza,” which he named for Eliza Doolittle, the heroine of George Bernard Shaw’s “Pygmalion,” a commoner who is “programmed” to speak and comport herself like a woman of the upper classes.

In the case of the digital Eliza, Weizenbaum designed it to interact with humans using natural language (rather than computer code), and in this case, to parody a psychotherapist by responding to its interlocutor by repeating back his or her own words, by employing stock phrases, or, as in:

"My boyfriend made me come here."
"Your boyfriend made you come here?"

"He says I'm depressed much of the time."
"I'm sorry to hear you are depressed."

To Weizenbaum’s surprise and distress, many people who interacted with Eliza thought they were conversing with an intelligent being (meaning that she passed what is known as the Turing test), and even found such “chatterbot” exchanges (as they are now called in computer lingo) helpful.

If the Nazis knew Eliza

He went on to fear the abuse of computers by, for example, fascistic governments vis-à-vis unsuspecting publics that fail to comprehend the limits of artificial intelligence. As he told the Boston Globe in a 1981 interview, “Since we do not now have any ways of making computers wise, we ought not now give computers tasks that demand wisdom.”

In his 1976 book “Computer Power and Human Reason,” Weizenbaum made his case regarding the dehumanizing dangers of computers. While he didn’t argue against their use in general, he did try to lower people’s expectations of them.

Many of Weizenbaum’s colleagues and peers responded skeptically to his criticisms, and saw him as something of a Luddite.

In 1996, Weizenbaum returned to Berlin. He spent the rest of his life there, writing and speaking about the dangers of computers, as he saw them. He died on March 5, 2008, at the age of 85.