Former Google boss says AI could change war in the same way as nuclear weapons

[midian182]

In context: With so many of today's headlines relating to AI in some form or another, it's easy to understand why people are calling 2023 the year that artificial intelligence really takes off. But, as with many technologies, there are plenty of potentially bad applications for AI alongside the good ones, including its impact on warfare. According to former Google CEO Eric Schmidt, artificial intelligence could have a similar effect on wars as the introduction of nuclear weapons.

In an interview with Wired, Schmidt, who in 2016 became a founding member of the government's Defense Innovation Board, talked about upgrading the US military with the most cutting-edge tech to take on China.

Some of Schmidt's conversation topics include a startup called Istari that uses machine learning and the metaverse to virtually assemble and test military vehicles such as tanks. He also talked about his vision for an internet of things-style network of inexpensive drones that could work seamlessly as one unit.

Schmidt left Alphabet's Board of Directors in 2019

Schmidt said the Pentagon should learn from the Russian invasion and how Ukraine has resisted its more powerful invader by moving quickly and adapting technology from the private sector, such as converting commercial drones into weapons and 3D-printing spare parts—though Starlink has put a stop to its SpaceX service being militarized. "Let's imagine we're going to build a better war-fighting system," Schmidt said. "We would just create a tech company."

But the real game changer could be artificial intelligence. "Einstein wrote a letter to Roosevelt in the 1930s saying that there is this new technology—nuclear weapons—that could change war, which it clearly did. I would argue that [AI-powered] autonomy and decentralized, distributed systems are that powerful," Schmidt said.

The former Google CEO and Alphabet chair has helped bring his vision of AI-powered warfare into the DOD over the last decade. Agency leaders now believe the technology will revolutionize military hardware, intelligence gathering, and backend software, helping the US in the arms race against China.

Comparing AI to nuclear weapons doesn't sound like a positive thing, even if Schmidt is making the comparison in terms of how it could change the way wars are fought, rather than having the potential for outright annihilation. Plenty of people believe AI can be used for good—Nvidia boss Jensen Huang called ChatGPT the greatest thing to ever happen to the computing industry—but with Lockheed Martin just announcing that an AI recently flew one of its jets for 17 hours, many remain concerned about the technology's applications when it comes to warfare.

Permalink to story.

https://www.techspot.com/news/97601-former-google-boss-ai-could-change-war-same.html

 

[NewSparrow]

I wouldn't be surprised if they call it Skynet within the next decade or so, scary times ahead.

 

[wiyosaya]

IMO, anyone at any level of industry or government that advocates handing over control of war to AI is certifiably insane.

Interestingly enough, there has been Sci-Fi imaginations of what that would be like:
https://www.imdb.com/title/tt0708414/?ref_=tt_urv
https://www.imdb.com/title/tt0064177/?ref_=nv_sr_srsg_0
As only two examples.

But, to quote Forrest Gump: "Stupid is as stupid does".

AI is ultimately computer code written by humans. It, in reality, has no sense of ethics, decency, or human values other than, in the broadest sense, those programmed by humans. We all know humans are fallible, and to think that no one would try to exploit such a programing opportunity is, IMO, simply to place too much faith in humanity that whom ever did the programming always acted in the manner that would ensure the best possible outcome for humanity.

I, for one, am not ready to place that kind of faith in a human programmer.

Further, I think we place far too much faith in people who have made lots of money as if they somehow have a lock on "genius" or "success" in all areas of human development. I see it as a by-product of society. Most people l think that such people, presumably "having it all," is a good thing that we all should aspire to. Sure having money could mean comfort, but below that facade, what drove these people?

 

[psycros]

The moment that we create an AI with not just the reasoning capacity but also the emotional intelligence to make its own decisions about war, it will either try to kill every human alive or refuse to fight at all.

 

[mbrowne5061]

<p>..though Starlink has put a stop to its <a href="https://www.techspot.com/news/97552-spacex-stops-ukraine-using-starlink-service-control-drones.html">SpaceX service being militarized</a>.</p>

No they didn't. They just moved the service over to its own dedicated satellites:

SpaceX reveals 'Starshield' satellite project for national security use

Starshield will leverage SpaceX's Starlink broadband technology for government customers.

www.space.com

 

[king cleanass]

This is a great idea.

The best part is that nothing could possibly go wrong.

 

[m4a4]

We trust war to humans. How could AI get any worse?

 

[yannus]

IMO, anyone at any level of industry or government that advocates handing over control of war to AI is certifiably insane.

I agree and I'd add that anyone who gives the monopoly of weapons to a government is also insane and we all did ( except for the US ).

 

[Theinsanegamer]

I agree and I'd add that anyone who gives the monopoly of weapons to a government is also insane and we all did ( except for the US ).

And in the last few years it's become REALLY apparent!

 

[toooooot]

How did nuclear weapons changed war?
Biggest nations still have small fights while owning nuclear weapons, example: India China.
It will make killing easier for sure.
I imagine that in the far future, there will be two highly sophisticated devices fighting with each other while soldiers waiting without engaging in combat because when one of these devices would lose, it would easily kill human soldiers.

 

[rsmith222]

I wouldn't be surprised if they call it Skynet within the next decade or so, scary times ahead.

Already happened.

The Google cellphone-drone-bomb technology that the US military has been using in the Middle East for a long time is called Skynet. People aren't very creative.

 

[rsmith222]

We trust war to humans. How could AI get any worse?

Take a look at some of the stable diffusion results.

 

[BuckarooBonzaii]

I'll be watching when our AI is stolen and sold to China or other Countries. Wasn't there a time when Israel sold U.S. weapons to China?

 

[yannus]

We trust war to humans. How could AI get any worse?

"Alexa tells 10-year-old girl to touch live plug with penny" --> War with AI could look either awkward either scary.

 

[OortCloud]

We need to round up all the politicians and CEO meglomaniacs and stick them all their f**king rockets and shoot them all at the sun. Can you imagine how much better the world would be without these morons.

 

[rsmith222]

The moment that we create an AI with not just the reasoning capacity but also the emotional intelligence to make its own decisions about war, it will either try to kill every human alive or refuse to fight at all.

Emotional intelligence? Emotions are heuristics, not full intelligence. A critical benefit of AI is that it would have the computational resources to not need to rely upon hacks like emotions.

The best hope humanity has is for AI to take over governance. Humans have clearly shown themselves to be utterly incapable of even the basics, like ensuring that train cars are inspected and repaired.

That humans in the 'world's greatest nation' can't even manage that, it says all that needs to be said about our ability to manage ourselves let alone the global ecosystem.