News for hackers. Stuff that matters.
Re:Google’s new AI ethics board is already falling apart by xiphias2 on Wednesday April 03, 2019 @08:10PM (#19568309)
From what I read about AI researchers and their optimism, I'm on the side of belief that it is the nature of intelligent life to destroy itself (https://en.m.wikipedia.org/wiki/Fermi_paradox?wprov=sfla1).

An ethics board would never be strong enough to stop AI development in the world if needed. We can't even limit CO2 output of the world with policy, and AI will be much more profitable than oil.

There's a strong agreement between most AI researchers that the biggest problem AI creates is the loss of jobs, and AGI is just too far to worry about, but I feel that at the same time they all hope to be the one who create the first AGI.

Re:Google’s new AI ethics board is already falling apart by est31 on Wednesday April 03, 2019 @09:09PM (#19568543)
There is a difference in the problems that AI might create in the future (destroying all humans) vs the problems it creates now. E.g. AI products had worse STT capabilties for people with a spanish accent or from black backgrounds than american english speakers. Other AIs had more errors with black people.

This board's mission is not to prevent an AGI to become self aware and destroy all humans as a natural conclusion of being self aware. Personally I think this scenario is pretty unlikely anyway.

Re:Google’s new AI ethics board is already falling apart by xiphias2 on Wednesday April 03, 2019 @09:47PM (#19568694)
I'm more worried about attack vs defence assymetry increasing in the world than a self aware AI destroying it.

You just need 1 crazy human and extremely strong assymetrical warfare. If you look at the cases when a human killed lots of people with a truck, and see that in 20 years with cheap self flying drones a crazy person can kill even more people, you can extrapolate this and get to a sad conclusion (at least I did that).

Re:Google’s new AI ethics board is already falling apart by 0x2AF31DA8 on Thursday April 04, 2019 @01:33AM (#19569582)
https://qntm.org/asteroids

> I believe that there is a threshold of power beyond which nobody can be trusted. Where, in fact, it is impossible for any entity to even theoretically demonstrate the track record of judgement, responsibility and infallibility that would be necessary.

Re:Google’s new AI ethics board is already falling apart by xiphias2 on Thursday April 04, 2019 @01:52AM (#19569654)
This was amazing read, thanks!

Also I believe it overestimates the power needed to kill all people.

How much time does it take for us to get to the point when an entity has live GPS coordinates of all people living in the world for example? Just 15 years ago this question sounded stupid.

Re:Google’s new AI ethics board is already falling apart by est31 on Thursday April 04, 2019 @02:36PM (#19575657)
That's indeed a far more likely scenario. There are plenty of crazy humans around right now. It sure is a big threat, but AI can also help to prevent damage. To stay with your truck example, Anis Amri for example would have killed far more had the truck not had a system that automatically engages brakes when there is an obstacle. Similarly, one could build strong AI systems to shoot down any drone attack. Bioweapons are more dangerous IMO because they are harder to defend from.
Re:Google’s new AI ethics board is already falling apart by gowld on Wednesday April 03, 2019 @09:51PM (#19568714)
There are so many ways we'll kill ourselves and crush our societies before AGI arrives. Pollution is just the start.

If Elon on Mars and Peter Thiel in his New Zealand compound end up getting an AGI after the rest of us are dead, it just doesn't matter to me.

Re:Google’s new AI ethics board is already falling apart by HNLurker2 on Thursday April 04, 2019 @12:19AM (#19569270)
You imply they are striving for immortality? Reminds me of a Peter Thiel interview.