Artificial intelligence does not pose a threat to humankind, Estonian professor Tammet says
Tanel Tammet, head of the Chair of Network Software of the Department of Computer Science at Tallinn University of Technology, said that prediction by an Oxford University scientist Nick Bostrom, according to which an artificial intelligence can soon become fatal to the humankind, belongs to the “field of scientology”, arguing that there have been no significant breakthroughs in the last 70 years and it is naive to think that it will happen in an equivalent period.
The job of Nick Bostrom, a researcher at Oxford University, is to explore whether artificial intelligence can spell the end of humankind and thus humans can become an extinction endangered species. Bostrom and several other researchers are of the opinion that computers and robots that outsmart humans will be produced already in this century, Finnish daily Helsingin Sanomat recently wrote.
According to Tammet, assessment of such a possibility depends primarily on what is meant by artificial intelligence. “For example, if we take such a very trivial thing that computers can add, multiply and play chess better, then this has been a fact for a long time already. There are several other things that they can do better.” More such things will definitely be added over a century, he said to Estonian daily Postimees.
Slow progress
However, artificial intelligence can be viewed otherwise, as the so-called social artificial intelligence formed as a result of human-like brain cells. “This exists already without computers. Computers certainly enhance it and the society as a whole becomes more powerful thanks to the addition of semi-artificial intelligence cells,” he said.
“If we now think that we take a computer and insert a lot of knowledge and the computer will then outsmart a human, then I believe that it is very unlikely that this will happen in a century,” Tammet said. Tammet explained that since progress in this area has been very slow in the last 70 years and there have been no significant breakthroughs, there is no reason to believe that this will happen in the next 70 years. “Maybe, but I do not believe that it is going to happen so soon.”
Humankind’s destruction unlikely
Bostrom and the researchers agreeing with him assume that highly intelligent machines can produce new generation machines with higher and higher intelligence. Finally, artificial intelligence will become super powerful and might consider humans superfluous and start destroying this species. Tammet said that this is very unlikely.
He brings an example of Protozoa and Metazoa from evolution - Protozoa were not destroyed when Metazoa came into being, they continue to coexist. “Based on this metaphor there is no reason to believe that if the society becomes more powerful, such single members of the society as humans would die out. This is very unlikely. They will, however, be united and organized in a different way,” Tammet said.
Tammet believes that even if artificial intelligence becomes more powerful, humans will learn how to coexist with it. “If we look at artificial organizations, such as ants or termites. Ants and termites have existed for a very long time and although humans are much more powerful than ants, it has never crossed our mind to kill all the ants. Such an idea has not occurred to us and there is no reason to believe that artificial intelligence will decide to kill all the ants and humans. This seems absurd,” Tammet said.
Restriction of the risks of artificial intelligence is absurd
According to Tammet, the whole Bostrom’s theory seems absurd as well, because he is a philosopher, not a researcher. Although more and more has been said about the threats of artificial intelligence in the last couple of years, it is mere scam in the professor’s opinion. “There are numerous organizations who act a bit like scientologists – they tell their threat story in abstract terms and then request money and say that they will save humankind from extinction.”
Bostrom points out various ways to prevent or restrict the threats of artificial intelligence, but Tammet says that this is impossible. “It is a totally naive idea. It is an absurd thought. It cannot be stopped or predicted in any circumstances.”
“We do not have to worry about the extinction of humankind. Not at least for this reason,” Tammet said, according to Tallinn University of Technology.
Editor: S. Tambur