Artificial Intelligence: Saviour, SkyNET, or ... video game nerd?

[Joey]

Wasn't sure where to put this, so figured I'd make a new thread.

First, we've got IBM and MIT, two of the most influential think-tanks in modern history, teaming up to take AI to the next level:
IBM team up on $240 million effort to rule the AI world

And now this:

What are your thoughts on AI? We've got some pretty brilliant minds that are terrified of it, going so far as to write an open letter on the potential perils of AI.

But then we've got Elon Musk revolutionizing Machine Learning like this.

Personally, I'm pretty excited to see how far this goes in our lifetimes.

[Nilanka]

I'm torn on the issue. Lots of brilliant minds telling us to tread lightly, while other brilliant minds tell us there's nothing to worry about (yet).

I'm not one to inhibit scientific/technological progress. It's necessary and inevitable. But I just hope we consider all likely outcomes when legislating this new endeavor.

[rocwell]

I'm torn on the issue. Lots of brilliant minds telling us to tread lightly, while other brilliant minds tell us there's nothing to worry about (yet).

I'm not one to inhibit scientific/technological progress. It's necessary and inevitable. But I just hope we consider all likely outcomes when legislating this new endeavor.

I think everyone agree that Al needs to be strictly regulated.

[Bendit]

I agree with Nilanka's sentiments. One needs to look at the other great "discovery and creation" of our time two of them in fact...nuclear and the internet. Both powerful forces for good....and unfortunately evil. Even "free speech" a wonderful societal gift which is/should be a human right is sometimes outright not permitted completely or is abused in using the permission many societies offer to foist segregation and racist principles for governance normally always affecting minorities. A good recent example is the expulsion of the entire Rohingya minority from Myanmar and supported it seems by a previous Nobel prize winner.

Sorry got carried away It's inevitable of course that research shall continue in the field but tenuous it shall be re it's use. Many aspects are being currently developed/used by military establishments. Hacking of the internet seems to be unmanageable so even if the "good guys" get a leg up on discovery and usage what happens re a hack into AI research secrets? Do we really believe that NK did not get a boost in their recent development of missile and nuke efforts/tests? I would bet they hacked into some of the western or even Pakistani or Indian military establishments for their secrets. The hacked of course would be loathe to divulge that they were compromised.

ps...I just remembered that medicine/biology (cloning and harvesting of organs) are also becoming issues with moral implications.

[Apollo]

Bendit, the key difference here is that nukes and the Internet aren't conscious. There are many risks with AI but the biggest is them forming a breakaway society and seeking to punish or exterminate their "captures". The tier below that is that AI will cause a huge wealth divide much larger than we see today. One proposed idea is to give each person an income regardless of whether they can work. The other idea is to tax the machines and their owners, make it so it's unfeasible to phase out all humans. I like the second option but I think the first is much more likely given how things are and how they've been heading without AI in the mix.

I think there needs to be a global charter signed by all nations agreeing on what nations are and aren't allowed to do with AI.

[Axel]

Skynet or the Matrix, either way, humanity is eventually doomed by AI. Personally I'm pulling for Matrix style destruction.

[Bendit]

Bendit, the key difference here is that nukes and the Internet aren't conscious. There are many risks with AI but the biggest is them forming a breakaway society and seeking to punish or exterminate their "captures". The tier below that is that AI will cause a huge wealth divide much larger than we see today. One proposed idea is to give each person an income regardless of whether they can work. The other idea is to tax the machines and their owners, make it so it's unfeasible to phase out all humans. I like the second option but I think the first is much more likely given how things are and how they've been heading without AI in the mix.

I think there needs to be a global charter signed by all nations agreeing on what nations are and aren't allowed to do with AI.

You make great points starting off (the conscious/unconscious) but end with something eerily similar to the nuke proliferation treaty. Treaties are fine until someone like a Kim (and I don't particularly want to dump on him) or others before him got away with taking a bite of the "forbidden" fruit. I had a bit of a giggle on the tax proposition considering we have a few posters who find evil in the imposition of any tax (= loss of freedom).

Humans as we know ourselves today...are fucked... with AI able to formulate a kind of evolutionary existence which theoretically would be exponentially faster than humans could ever naturally achieve...no?

I admit I am just rambling here so feel free to kick me in the nutz on this.

[Nilanka]

[Maury]

[Bendit]

Thought I'd place this interesting piece here...

A New Way for Machines to See, Taking Shape in Toronto

https://nyti.ms/2ibYkXB

[Bendit]

I trust this vid is not "doctored".....but still!! I believe it's sped up some.

[Joey]

I trust this vid is not "doctored".....but still!!

So satisfying to watch.

[Bendit]

So satisfying to watch.

Yeah...the precision! Couldn't see this working in north America...the Japanese are disciplined and sticklers. Can't imagine someone "asleep at the switch" or not "greasing the skids"...lol. Heard they are now testing our intrepid TTC workers for "peds" and a few were bounced on the first go around.

[rocwell]

Checkmate humanity: In four hours, a robot taught itself chess, then beat a grandmaster with moves never devised in the game's 1,500-year history and the implications are terrifying

http://www.dailymail.co.uk/sciencete...ves-hours.html

[JimiCliff]

Checkmate humanity: In four hours, a robot taught itself chess, then beat a grandmaster with moves never devised in the game's 1,500-year history and the implications are terrifying

http://www.dailymail.co.uk/sciencete...ves-hours.html

Not that it really matters in the grand scheme of things, but the robot didn't beat a grandmaster, it beat another robot.

It's astonishing how many journalists/news sources just flat-out dgaf about reporting things accurately.