With long hair, round glasses and a leopard-printed wide-brimmed hat, the Brazilian-born scientist caused a stir this week at the Web Summit conference in Rio de Janeiro, introducing Grace, a robotic nurse designed to care for the elderly in nursing homes.
It is to him that we owe the name of "Artificial General Intelligence" (AGI), this major technology with human cognitive abilities that could be developed "within a few years". He is at the origin of the SingularityNET platform, intended for the decentralized and democratic development of AI.
QUESTION: What is missing to develop an artificial intelligence system with human cognitive capabilities?
ANSWER: "If we want machines that are really as intelligent as people, and able to react quickly to the unexpected, they should be able to do much more than they are programmed to do. And we're not there yet. But there is reason to believe that this could happen, not in decades, but in a few years."
Q: What do you think of the debate around ChatGPT? Should we take a six-month pause in research, as some are advocating?
Mathematician Ben Goertzel and artificial intelligence robot Desdemona, at the Web Summit Conference, May 3, 2023 in Rio de Janeiro, Brazil © MAURO PIMENTEL / AFP
A: "I don't think we need to pause, because it's not a dangerous artificial intelligence (...) It cannot make complex, multi-step reasoning, like scientists, or invent new things outside of the data received. It cannot develop strategies to deal with new situations, such as a pandemic, for example.
Some people say that we need to stop research because this kind of system contributes to misinformation. But then, should we ban the internet? (...) I think we need to live in a free society and, just as the internet shouldn't be banned, we shouldn't ban it (ChatGPT) either."
Q: Should we see the possibility of AI replacing jobs as a threat?
A: "With the upcoming arrival of new systems of the same type as ChatGPT, I think that 80% of the jobs held by humans could become obsolete. But I don't see it as a threat, but as an advantage. That is a good thing. People will find better things to do than work for a living. Virtually all administrative tasks can be automated."
"The problem will be the transition period, when artificial intelligences will start to make obsolete one job after another (...) I don't know how we're going to solve the social problems caused."
Q: What contributions can robots make to society today, and what could they do in the future, with AGI?
A: "They can do a lot of good things, like Grace, the robot nurse. In the United States, a large number of elderly people feel lonely in nursing homes. Even if they are fed, receive medical care and watch TV, it is insufficient in terms of social and emotional support.
If we bring humanoid robots into these spaces that answer their questions, listen to their stories, help them call their children or shop online, it improves their lives."
The artificial intelligence robot Desdemona, at the Web Summit Conference, on May 3, 2023 in Rio de Janeiro, Brazil © MAURO PIMENTEL / AFP
"In this case, we do not cut jobs, because there are not enough candidates for nursing or nursing assistant positions. Education could also be an incredible market for humanoid robots, as could housework."
Q: How do you regulate AI so that it has a positive impact on people's lives?
A: "Governance should be participatory, involve the people, in a way. And it's technically possible. The problem is that the companies that fund most AI research are not interested in the common good. What they want is to make as much money as possible for their shareholders."
Interview by Eugenia LOGIURATTO
© 2023 AFP