AI might replace 80% of occupations in a few years.

In the coming years, artificial intelligence could replace 80% of human jobs, but this is a positive thing, says Ben Goertzel, an American-Brazilian AI expert.

Goertzel, a 56-year-old mathematician, cognitive scientist, and renowned robot-designer, is the founder and CEO of SingularityNET, a research organization he founded to create “Artificial General Intelligence,” or AGI — artificial intelligence with human cognitive abilities.

Goertzel, with his long hair and leopard-print cowboy hat, was in provocateur mode last week at Web Summit in Rio de Janeiro, the largest technology conference in the world, where he told AFP in an interview that AGI is only a few years away and criticized recent efforts to restrict artificial intelligence research.

Intelligent as humans?

How close are we to developing artificial intelligence with human-level cognitive capabilities?

“If we want machines to be as intelligent as humans and as adaptable in the face of the unknown, they must be able to make large advances beyond their training and programming. And we’re still not there. However, I believe we are years rather than decades away from achieving this.”

AI risk

What are your thoughts on the debate surrounding AI such as ChatGPT and its associated risks? Should there be a six-month research hiatus, as proposed by some?

“I do not believe that we should halt it because it resembles a dangerous superhuman AI… These AI systems are extremely intriguing, but they are incapable of achieving human-level general intelligence because they cannot engage in the complex, multi-stage reasoning required for scientific inquiry. They are unable to conceive novel concepts outside the scope of their training data.

“They can also disseminate false information, and some argue that we should halt their use because of this. That seems very odd to me. Why hasn’t the internet been banned? Internet does precisely this. It makes significantly more information accessible. And it promotes misinformation and nonsense.

“I believe we ought to have a liberated society. And just as the internet should not be prohibited, neither should this be.”

Danger to employment

Isn’t their potential to displace human workers a threat?

“Without an AGI, you could probably eliminate up to 80 percent of the tasks that people perform, according to my estimation. Not precisely with ChatGPT as a product. However, systems of this character will emerge within the next few years.

“I do not believe it poses a hazard. I consider it a benefit. People can discover more enjoyable activities than working for a living… Practically every paper-based task should be automatable.

“The problem I see is in the interim, when AIs are eliminating human jobs one by one… I do not know how to resolve all social problems.”

AI positives

What can machines do for society now, and what will they be able to do if AGI is realized?

“You can accomplish much good with AI.

“We displayed Grace (a automated nurse) at Web Summit Rio. A large number of elderly Americans reside in old folks’ homes alone. And they’re not terrible in terms of physical condition — you have access to medical care, sustenance, and a large-screen television — but they lack emotional and social support. Therefore, if you inject it with humanoid robots that can answer your queries, listen to your stories, help you call your children or place an online order, you are improving people’s lives. As your AGI increases, they will become even greater companions.

“In this circumstance, you are not eliminating jobs for humans. Because there are not enough individuals interested in nursing and nursing assistant positions.

In addition to domestic assistance, I believe education will be a tremendous market for humanoid machines.

Regulation

Which regulations are necessary for AI to have a positive impact?

“You need society to cultivate these AIs so that they do good things. And the governance of AIs should involve the population in some way. All of this is technically feasible. The issue is that the companies funding the majority of AI research are not concerned with doing decent work. They are concerned with maximizing shareholder value.”