Artificial Intelligence Will Do What We Ask. That’s a Problem.

Binary_Bark

Binary_Bark

Forging
Joined
Feb 24, 2016
Messages
41,689
Reaction score
23,614
Location
Midgard
The danger of having artificially intelligent machines do our bidding is that we might not be careful enough about what we wish for. The lines of code that animate these machines will inevitably lack nuance, forget to spell out caveats, and end up giving AI systems goals and incentives that don’t align with our true preferences.

A now-classic thought experiment illustrating this problem was posed by the Oxford philosopher Nick Bostrom in 2003. Bostrom imagined a superintelligent robot, programmed with the seemingly innocuous goal of manufacturing paper clips. The robot eventually turns the whole world into a giant paper clip factory.
Click to expand...
More At: https://www.quantamagazine.org/artificial-intelligence-will-do-what-we-ask-thats-a-problem-20200130/
 
Sorry. TLDR. So I didn't read the whole article.
I get the point though, but this is why it's named 'artificial' intelligence.
Humanity will always be smarter.
Terminator and iRobot won't be reality, even with quantum advancement.
The kill switch is initiated by humans, not the Matrix.
 
Daruk said:
The trouble is, what if humans decide it's the lesser of two evils (letting a machine make high risk decisions) - e.g. you can't blame me for a big political decision - the data all pointed to the ultimate decision taken. There's AI and AGI and while it's debatable whether we'll achieve AGI, it's when people start working things to be fully autonomous you might find were a) too many kill switches away and b) built in too much redundancy and c) changed policy to require too many hoops to jump through to hit that 'kill switch'.
Click to expand...
It's a bit early in the morning for me to focus, but I definitely don't disagree with your statement. There are many variables to take into account.
 
Ponderer said:
AI is just an algorithm (a particular set of instructions).
AI is a misnomer - the algorithm is not "artificial", nor is it "intelligent".
Click to expand...

Highly simplistic view of the subject. There are interesting areas of AI, like emergent behaviour, where complex behaviour emerges not from a set of instructions telling each part what to do, but by the individual interaction
 
|tera| said:
Sorry. TLDR. So I didn't read the whole article.
I get the point though, but this is why it's named 'artificial' intelligence.
Humanity will always be smarter.
Terminator and iRobot won't be reality, even with quantum advancement.
The kill switch is initiated by humans, not the Matrix.
Click to expand...
Okay, you keep telling yourself that then :)
 
vic777 said:
Highly simplistic view of the subject. There are interesting areas of AI, like emergent behaviour, where complex behaviour emerges not from a set of instructions telling each part what to do, but by the individual interaction
Click to expand...
Is this so called "emergent behaviour" part of the algorithm or not.
 
AI!=Machine Learning

I've yet to see anything resembling how biological neurons work on any sort of meaningful scale.

Using a series of PID loops daisy chained to make "neurons" is not the same thing, especially when you tune it to get results you want.

Please stahp.

For goodness sake, it is 2020 and I cant get a mac to talk to a new printer at the same times as a PC.
 
IF AI were to to achieve *whatever the hell you think should go here, goes here*.

Please explain why certain problems i.e: The Whitehead Problem is still unsolvable using Quantum Computing/ a Quantum Computer?
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X