2023-12-02 18:35:00
Bletchley Park, a manor house which combines Victorian Gothic, Tudor style, Dutch Baroque and… a fascinating past: it is here, in the county of Buckinghamshire, in the south-east of England, that Alan Turing and his band of brilliant mathematicians deciphered the Nazi Enigma communications system during World War II.
It was also there that, eighty years later, on November 1 and 2, the British Prime Minister, Rishi Sunak, the Vice-President of the United States, Kamala Harris, and even the President of the European Commission, Ursula von der Leyen, to discuss the future of artificial intelligence. In the aisles of this AI Safety Summit, we meet the best of researchers. Like Yoshua Bengio, who contributed to the development of the beginnings of structuring supervised learning, that is to say artificial neurons which, guided, are capable of recognizing characters, shapes or colors.
“We must prepare for dangerous uses of AI”
Born in Paris and settled in Canada, he was at the origin, in the mid-1990s, of generative adversarial networks, or generative adversarial networks: miniprograms that make the computer creative, in particular by inventing images. Usually calm, he appears a little worried. During the conference, he explains to Point that “ it is absolutely necessary to prepare for dangerous uses of AI “. An example ? “ This may involve making it easier for evildoers to develop chemical and biological weapons or to carry out cyberattacks, as well as to launch disinformation campaigns. »
And when asked about open source, an open approach that allows a community of researchers to continually fix bugs, he cautions: “ We do not want to hand criminals, terrorists and rogue states dangerous capabilities that are difficult to access today. » For him, the security requirement must be comparable to that required of companies that manufacture bridges, planes or medicines.
On the last day of the conference, Elon Musk, lawyer for the Neuralink project, which aims to graft a chip into the human brain in order to allow the latter, among other things, to resist the progress of artificial intelligence, goes so far as to predicting in front of a stunned Rishi Sunak the end of human work. This replacement of all our jobs is a theory that Brad Smith does not subscribe to. “ I do not believe it one second. I will always prefer a classroom where the teacher is a human being and not a computer and, if I go to the hospital tomorrow, I would rather talk to a doctor than to a machine », replies the president of Microsoft to Point, before saying: “ Intelligence quotient is not everything, we need empathy… Humans must remain in charge. »§
“I believe we are witnessing the emergence of the most disruptive force in human history. For the first time, we have here something more intelligent than the most intelligent of humans. It’s hard to say when that’s going to happen, but we’ll get to a point where there’s no need to work anymore. […] One of the challenges of the future will be to find meaning in our lives with this kind of magical genie at our disposal who can do whatever we want. »Elon Musk, during the AI Safety Summit, in Bletchley Park (England), November 1, 2023.
“It’s so incredible to think that we’re going to experience Time.news*. This will probably be the most important event in world history and it will happen in our lifetime. » Adam D’Angelo, on X, August 20, 2023.
” TRUE. I hope she will be kind to us humans. »Elon Musk, on X, August 20, 2023.
“The security requirement must be comparable to that required of companies that manufacture bridges, planes or medicines. »Yoshua Bengio, professor at the University of Montreal and 2018 co-recipient of the Turing Prize, at “Point”, November 2, 2023.
#Artificial #intelligence #work