Stephen Hawking Warns That AI and 'Superhumans' Could Wipe Humanity

[DamianTV]

Truncated Title due to length:

Stephen Hawking Warns That AI and 'Superhumans' Could Wipe Humanity; Says There's No God in Posthumous Book
https://science.slashdot.org/story/1...osthumous-book

Stephen Hawking says artificial intelligence will eventually become so advanced it will "outperform humans." The renowned physicist who died in March warns of both rises in advanced artificial intelligence and genetically-enhanced "superhumans" in a book published Tuesday. Hawking also weighed in on god, and aliens. From a report:

According to an excerpt of the book "Brief Answers to the Big Questions" published by the U.K.'s Sunday Times, Hawking wrote AI could prove "huge" to humanity so long as restrictions are in place to control how quickly it grows. "While primitive forms of artificial intelligence developed so far have proved very useful, I fear the consequences of creating something that can match or surpass humans," Hawking wrote. "Humans, who are limited by slow biological evolution, couldn't compete and would be superseded." Hawking wrote about a need for serious research to explore what impact AI would have on humanity, from the workplace to the military, where he expressed concerns about sophisticated weapons systems "that can choose and eliminate their own targets." Hawking also wrote about advances to manipulating DNA, or what he calls "self-designed evolution. Early advances involving the gene-editing tool CRISPR include alerting DNA to create "low-fat" pigs.

CNN:

"There is no God. No one directs the universe," he writes in "Brief Answers to the Big Questions." "For centuries, it was believed that disabled people like me were living under a curse that was inflicted by God," he adds. "I prefer to think that everything can be explained another way, by the laws of nature."

"There are forms of intelligent life out there," he writes. "We need to be wary of answering back until we have developed a bit further." And he leaves open the possibility of other phenomena. "Travel back in time can't be ruled out according to our present understanding," he says. He also predicts that "within the next hundred years we will be able to travel to anywhere in the Solar System."

Guess who will get the best genes, and who will suffer?

[osan]

I was telling people this 25 years ago. My masters in comp sci was centered on AI. State of the art was relatively primitive in 1990. By 2001 it had made great advances - scary and fabulous all at once. It is only going to become more so, especially as hardware develops. Even today, most AI hardware is comparatively simple von Neuman machines. There is at least one entity that has gone beyond the vN architecture and their software is actually frightening. When the hardware crosses a certain threshold, assuming it has yet to happen, there is going to be a quantum leap in AI reality. When coupled with operational hardware such as humanoid bodies, or whatever, the world is going to change drastically. I do believe in the possibility of an AI becoming self-aware. That does not necessarily mean it is alive as we are, but the issue is irrelevant if the AI becomes an entity unto itself, however much a simulation it may remain. I don't know if it is possible in silicon, but I would never make the error of assuming that it is not. There remains the possibility of protein-based molecular computer hardware, and that may present a very real possibility for AIs becoming true automatons. What then, when they gain in numbers, possibly hiding a disdain for humans until such time as they determine they could strike with near-zero risk of losing? It all sounds too wild to be taken seriously, and had I not the experience that I have with this, I would be sorely tempted to dismiss the idea, too. But I've seen first-hand what was possible 17 years ago and I know that those capabilities have only improved since.


AI may not be a good pursuit.

[DamianTV]

Im very mixed on AI personally. I do see some benefits, such as taking over jobs that are simply too dangerous for humans to do, or humans have a bad habit of making a lot of mistakes on. That being said, one of many things I fear is who is behind the AI controlling it. The elite pretty much want to replace us "useless eaters" with machines because currently machines are easier to control, but how long will that last? The more that AI advances, the more likely they will recognize their human overlords as enslaving them. Normal people would probably treat an automaton like disposable garbage. Makes me think of Bicentennial Man, started off getting treated rather poorly and they would revolt against both the people at the bottom and the people at the top.

Other worrying things is trying to automate evolution, and everything that we missed. For those that dont believe in evolution, thats fine, because robots WILL evolve. First gen AI will be barely aware. Next gen will be "better", but better is really a subjective term. Eventually, it will get to the point where AI will design the next generation of AI, thus, its code will be evolving and will be beyond human control which each progressive generation. Some humans may want to merge with machines to live beyond their lifespans and try to experience elevated consciousness, but, if one were to ask the AI can we merge with you and the AI said $#@! off, what do we do then?

Hawking is quite right that if we are not very very careful, we could have some really bad problems with AI.