Why this new videogame AI is so incredibly controversial

Ooh, the military applications for this seem "fruitful"

huge+(1).jpg


Above is the "ATLAS" prototype by Boston Dynamics (funded by DARPA) now if they can get this AI adapted to work with this...hmmm...
 
I felt so bad for that poor robot when it was trying to pick up the box but kept getting messed around:cry:

Looks like this tech has made some real progress though.

It still has its flaws, but they are working towards getting it close to the prototype photo. It will be (if they get everything right) a +/- 2m tall bipedal robot weighing ~300kg that can navigate rough terrain and fire a weapon.
 
"which rewarded it for killing more people, the fear is that if ported into the real world, it wouldn’t be satisfied with a single kill". These things don't have emotions. Rewarded means that it just puts a larger weight on things learnt to accomplish that task. That is what reward learning is. Goddammit
 
“Given that it was trained via deep reinforcement learning which rewarded it for killing more people, the fear is that if ported into the real world,
That is such horsesht.

You can apply deep learning to anything, but to say that this can be ported is like saying a computer chess game is a dangerous cause it could learn to invade country X.
 
It's funny that this bot AI for an FPS game turned out to be the controversial one. We've had adaptive/learning AIs in shooters for ages, but now this one will result in The Matrix / Judgement Day.

I also love how delightfully self-aware they were in naming their bot Arnold.

Hasta la vista, baby.
 
Back
Top