Max Tegmark has warned several times about the risks of artificial intelligence. This week, he, along with several other researchers, politicians and business leaders, wrote under a statement from the Center for AI Safety in which they singled out AI as an existential threat similar to pandemics and nuclear war.
IN SVT’s 30 minutes Tegmark elaborates on what he means by AI being able to exterminate humanity.
He draws parallels to animal species that have recently become extinct and believes that this happened because humans were the most intelligent creature.
– What we are warning about now is that if we humans lose control of our society to machines that are much smarter than us, then things can go just as bad, he says in the program.
50 percent risk
We cannot know how this would happen and, according to Tegmark, it is because a less intelligent species cannot imagine how it will happen when a more intelligent species exterminates them.
– Exactly how it would happen is pure speculation, but that it will happen is very likely, I would say more than a 50 percent risk, he says.
When Tegmark is asked how he can be sure that the risk is so high, he draws parallels to how it looked in history when a new, smarter species appeared.
– Pretty much every time it has gone quite badly for those who used to be the smartest before, he says.
Need regulation
Politicians need to take the threat from AI seriously, he says, and calls for tougher regulation of companies, similar to how it looks in the pharmaceutical industry.
For example, a pharmaceutical company may not start selling its medicine before it has undergone thorough checks by the Swedish Medicines Agency.
– It is the responsibility of the companies to prove that what they sell does not have major risks before they get permission, today it is just the opposite, it is the wild west.