The ancestor of ChatGPT is a lady, and he or she is 58 years outdated | EUROtoday
Get real time updates directly on you device, subscribe now.
Pto jot down your emails, set up your holidays, draw up your buying record or create web sites… Since November 30, 2022 – the date on which the American firm OpenAI made ChatGPT out there to most people – researchers, entrepreneurs and customers have continued to to take a position on this societal revolution as AI (synthetic intelligence) and chatbots achieve floor in our every day lives… To the purpose of generally forgetting that these conversational robots should not new.
ELIZA, ChatGPT's oldest grandmother, is 58 years outdated. And, if it didn’t attain the 200 million customers of its well-known inheritor, it however marked a decisive step within the historical past of computing and new applied sciences.
ELIZA, the primary chatbot
This very first conversational robotic was born in 1966, in a laboratory at MIT (Massachusetts Institute of Technology), within the United States. Its creator? Joseph Weizenbaum, a Jewish German-American pc scientist, born January 8, 1923 (and died in 2008), in Berlin, who emigrated to the United States along with his mother and father in 1936 to flee Nazi Germany.
After good research in arithmetic – interrupted for a time to function a meteorologist within the military in the course of the Second World War – this pioneer first labored at General Electric, the place he co-developed the primary pc system devoted to banking operations. He additionally created the “SLIP” programming language, devoted to record processing, which can permit him to put the foundations of the ELIZA program.
Three years after becoming a member of MIT, he launched his conversational robotic. The precept: put this chatbot within the footwear of a psychotherapist in whom customers can confide. By figuring out the important thing phrases of the “scripts” despatched, ELIZA then responds just by reformulating every assertion submitted within the type of a query, giving the impression that she understands her interlocutor and that she is exhibiting empathy.
ELIZA is a program that makes pure language dialog attainable with a pc.Joseph Weizenbaum
In the scientific article that he printed to current his program, Joseph Weizenbaum needed to chop quick fanciful interpretations that may give the machine capabilities that it doesn’t have. “ELIZA is a program that makes natural language conversation possible with a computer,” he writes. They say that to elucidate is to justify. Nowhere is that this maxim extra nicely revered than within the subject of pc programming and synthetic intelligence. Indeed, machines behave in marvelous methods, typically to the purpose of dazzling even probably the most skilled observer. But as soon as a selected program is unmasked, its inside workings defined, its magic fades; it seems to be a easy assortment of procedures, every of which is totally comprehensible. The observer then says to himself: I may have written that. The latter then strikes this system in query from the “smart” product shelf to that reserved for curiosities. »
The ELIZA impact, or anthropomorphism utilized to computer systems
Why did you title your program ELIZA? The pc scientist refers back to the fictional character Eliza Doolittle, within the play Pygmalion (1913) by George Bernard Shaw, additionally identified for its musical adaptation, My Fair Lady (1956). Like the flower vendor who speaks Cockney, well-liked London slang, and takes classes to grow to be a spokesperson for the excessive society elite, this pc program “can be improved by its users, just as its linguistic capabilities can be continually improved by a “professor”, specifies Joseph Weizenbaum. Like Pygmalion’s Eliza, she can be made even more civilized.”
However, as soon as his “chatbot” was available, the scientist was surprised to see how difficult it was for some users not to humanize this program. Some people even became dependent on their relationship with this virtual agent. This gave rise to a concept, the ELIZA effect, which describes the tendency to unconsciously equate the behavior of a computer with that of a human.
Starting with the secretary of the computer scientist who, one day when she was talking with the robot, asked Pr Weizenbaum, who was checking the effectiveness of his creation in action, left the room. “I believe this anecdote testifies to the success with which the program maintains the illusion of understanding,” he joked at the time.
Joseph Weizenbaum's fight against the “artificial intelligentsia”
But the fantasies and excesses surrounding the use of artificial intelligence quickly ceased to make this professor, who has taught courses at Stanford, Harvard, Berlin and Hamburg, laugh. To such a degree that “after having developed ELIZA, he will spend the rest of his life fighting what he himself called the “artificial intelligentsia”, these technological elites convinced that human beings and machines are interchangeable,” explains Olivier Tesquet, journalist in charge of digital issues at Teleramain the preface of the book Artificial intelligence: a (r)evolution?.
In 1976, Joseph Weizenbaum printed the essay Computer energy and human cause: from judgment to calculation (Computer energy and human cause – from judgment to calculation), humanist critique of synthetic intelligence, and particularly of methods that substitute automated decision-making for the human thoughts. He insists that delegating our selections to machines is just not solely mistaken, but in addition harmful, as a result of they’re merely incapable of knowledge and compassion.
“It raised questions early on about the kinds of relationships we want to have with machines,” he instructed the New York Times Sherry Turkle, professor at MIT, with whom she taught programs on the social implications of know-how.
If this criticism has pushed a wedge between this researcher and different members of the unreal intelligence group, the one who returned to Berlin to spend the final twelve years of his life is “still unjustly forgotten in the pantheon of the founding fathers of intelligence artificial”, concludes Olivier Tesquet, in a context the place “the resurrection of the ideas of the German-American computer scientist nevertheless sounds like an imperative necessity”.
https://www.lepoint.fr/eureka/l-ancetre-de-chatgpt-est-une-femme-et-elle-a-58-ans-15-05-2024-2560237_4706.php