ban on killer robots

[ian in lancs]

https://www.theguardian.com/technology/2017/aug/20...

Bit late now!

[xjay1337]

That means autonomous cars will be banned then?

[Murph7355]

ian in lancs said:

...
Bit late now!

Exactly.

AI was a module on my MSc in 1993, and one of the major parts of that was around the philosophical side rather than the technical (which was relatively easy).

[Gecko1978]

Its a real issue though as rich nations could build autonomous soldiers and use them to attack poorer nations and also abuse there own people. Robots turning on us is perhaps least of our worries. The technology is not there yet but in 20 years time I could see a tank being fully unmanned an able to operate on the battle field etc after all if a drone can then its only a few steps away. I doubt we will have humanoid soldiers as it might be sub optimal design but perhaps multiple units that could combine to climb stairs etc and why duck for cover when your made of metal etc just take the hits.

[98elise]

I've always thought it won't be long before Boston Dynamics robots are on the battlefield.

We've had autonomous defensive weapons since the 70's.

[s1962a]

Do we have to build mean, powerful, fighting robots to make sure other nations don't get them?

[Murph7355]

Gecko1978 said:

Its a real issue though ...

Technological advancement has always been a real issue when not fully thought through (and I'd say it rarely is).

Risk/reward balance.

Key thing is to be the first/best at it

There's a degree of irony with someone like Musk wanting it banned for a specific type of use. Not a prayer.

[FurtiveFreddy]

Oh great. I've just finished building mine and now I see this

[jurbie]

As we are never likely to eradicate warfare surely having robot armies fighting each other is better than human armies?

[Jonesy23]

So is this an argument based on a genuine knowledge of the risks ('cos some of their 'experts' really aren't), a nice fluffy moral argument, or something cooked up by some of the commercial players in that group to push some well funded competitors out of the game?

Musk rarely does something that isn't ultimately about publicity or money and this is an area he is trying to break into so I suspect altruism has nothing to do with his new concern about the so called dangers of AI.

And it's funny how the same concern doesn't stretch to handing potential moral decisions to a couple of tonnes of hurtling metal if he's selling it.

[Murph7355]

jurbie said:

As we are never likely to eradicate warfare surely having robot armies fighting each other is better than human armies?

A robot can't "own" territory gained though. The robots wouldn't stop at destroying other robots.

As someone mentioned above, we've had autonomous weapons for decades. There's no way of legislating against them IMO. And if we start now, are we also going to retrospectively ban automatic targeting systems etc? Or things like the Missile Defence Agency?

I've thought for a while that an autonomous (semi or fully) combat aircraft would be a massive advantage as I'm sure the person sat in the cockpit is currently the biggest limiting factor from a physical point of view.