Mark my words AI is much more dangerous than Nukes- Elon Musk

[drcoastline]

I am sure this will end up in Parf but I'll start this here.

[rusnak]

I wonder if he thinks its a very serious danger to the public

[drcoastline]

Originally Posted by rusnak

I wonder if he thinks its a very serious danger to the public

Yes he does. He thinks it should be regulated because of the danger.

[911_Dude]

Musk has been saying this for a long time. But the video really does not offer anything of value, except Musk saying AI scares him. If you want a good read on what the issue is with AI read this (its long, but an easy read) https://waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html





.

[Tervuren]

One of the biggest threats AI poses is control of information feeds for evil ends.

AI can be used to identify would be trouble makers and feed them the information that leads to violence. It is a weapon of war, and I would not be surprised if it is being used the past few years.

The insidious thing is people can be visiting the same URL's for informationn, but the AI can feed different people different information. They won't even realize they are being played.

The AI doesn't have to launch nukes to destroy humans, the AI just has to currate their news and social media. At which point, the AI can get people to destroy other people. Yes, it is a threat, and being aware of it is a major part of fighting it.

911 Dude, I struggle with that article as hunter gatherer's are still around today, as are basic agricultural villages.
These people aren't just dieing from shock if a land rover drives by for the first time.

[Norm K]

Humans are so bad at writing software, what makes us think that the first truly sentient AI won't be insane?

Once we discover it, of course, we'll set teams of software engineers to re-writing the code, something that shouldn't take more than a few weeks. Meanwhile, it re-wrote itself in a few seconds, then faster, and faster yet....

We're domed!

_

[Nostril Cheese]

[911_Dude]

The problem is, the most money and research into AI is probably with military and Wall Street. I dont want either one accidently creating "The Singularity".

[island911]

The only reason Musk says all of that is because his AI told him. - now what?

[island911]

AI is but a tool.

The problem sits with the tool that a human may build. - what is 'the game' that the AI is tasked with?

[Eric Hahl]

[Tervuren]

Originally Posted by Eric Hahl

That is the way it goes in major Pc games these days, at least, it feels like.

[GH85Carrera]

Originally Posted by Nostril Cheese

One of the misunderstood plot points in 2001 is HAL did not just take over and try to eliminate the humans for no reason. He had deep planted instructions to get the ship to the monolith orbiting Jupiter at any and all costs. The human programmers essentially told HAL that if the humans get in the way of the mission they are expendable. It is explained in 2010 A Space Odyssey and in the book of the same name.

Arthur C. Clarke based the move 34 year in the future, and here we are 20 years after 2001 and we still are far from a conscious intelligent self aware computer, or sending humans to Jupiter.

[Porsche-O-Phile]

Skynet. It’s coming...

[Steve Carlton]

2010: The Year We Make Contact is one of my favorite movies. Good characters and good story.

The Terminator documentaries really illustrate the widespread danger of AI.

[Alan A]

Originally Posted by 911_Dude

The problem is, the most money and research into AI is probably with military and Wall Street. I dont want either one accidently creating "The Singularity".

china rather than wall street i reckon.

if you (all) want a free read (assuming you have kindle unlimited) i was given a copy of this just after it came out as a conference takehome.
it's now available on kindle unlimited.

https://www.amazon.com/AI-Superpowers-China-Silicon-Valley-ebook/dp/B0795DNWCF/ref=sr_1_1?dchild=1&keywords=AI+Superpowers%3A+China%2C+Silicon+Valley %2C+and+the+New+World+Order&qid=1612807035&sr=8-1

[GH85Carrera]

The assumption in movies is always that the IA will be malevolent, and not benevolent. I guess if the computers were like in Star Trek and just were helpful and wanted to help better life for mankind there is no conflict and and hence no plot.

Assuming the IA can rewrite it's own code and get smarter and more capable the real challenge will be to make a robot that can make things and manipulate the world around itself like humans can with a thumb and fingers and muscles. Why it will feel it needs to fight and kill humans is just the plot point for the movies. For sure we need to be careful with real cognitive AI.

And until AI is so smart it can change the laws of thermodynamics and physics Terminators will never be a real threat. Certainly a malevolent, AI can engineer a virus deadlier to humans than anything on Earth now. We can just pull the plugs and shut down all electricity generation and get medieval on it's ass. No AI can take over my old 1980s cars. Cars from the 60s and before are safe from AI hacking.

[Tervuren]

You're thinking in terms of conventional warfare between AI and humans.

The AI doesn't have to use conventional warfare, it just has to turn us against ourselves.
And when it does it, it will be smart enough we won't even know to pull the plug.
Humans are irrational first, and if rational at all, it is generally an afterthought to justify your actions to other people. Only a small amount of human action is based on rationality.
Which makes us excellent marks for an AI to exploit.

[Crowbob]

Originally Posted by GH85Carrera

The assumption in movies is always that the IA will be malevolent, and not benevolent. I guess if the computers were like in Star Trek and just were helpful and wanted to help better life for mankind there is no conflict and and hence no plot.

Assuming the IA can rewrite it's own code and get smarter and more capable the real challenge will be to make a robot that can make things and manipulate the world around itself like humans can with a thumb and fingers and muscles. Why it will feel it needs to fight and kill humans is just the plot point for the movies. For sure we need to be careful with real cognitive AI.

And until AI is so smart it can change the laws of thermodynamics and physics Terminators will never be a real threat. Certainly a malevolent, AI can engineer a virus deadlier to humans than anything on Earth now. We can just pull the plugs and shut down all electricity generation and get medieval on it's ass. No AI can take over my old 1980s cars. Cars from the 60s and before are safe from AI hacking.

It won't have to take over cars. It will take over the refineries.

[Wetwork]

Originally Posted by 911_Dude

Musk has been saying this for a long time. But the video really does not offer anything of value, except Musk saying AI scares him. If you want a good read on what the issue is with AI read this (its long, but an easy read) https://waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html

.

Just got done reading the whole thing....Just great, all I thought I'd have to worry about as a adult was quicksand, and killer bee's. Good read though, if the big heads are worried about it I hope folks actually listen.-WW