5 ms·

As opposed to tired, stressed and easily overwhelmed humans making tactical and strategic decisions? I share your concern but I don't think its impossible for a

by jamii 16y ago

As opposed to tired, stressed and easily overwhelmed humans making tactical and strategic decisions? I share your concern but I don't think its impossible for an AI to get to the point where it makes less mistakes than a human.

khafra 16y ago
The stereotypical danger isn't that they'll make fewer mistakes than a human, it's that the mistakes they do make will be much bigger--continually offloading decisionmaking and decision enforcement to AI without a coherent theory of friendliness is like using Martingale betting: You don't lose often, but it only takes one to wipe you out.
- jimbokun 16y ago
  Much Paper Clip AI potential. http://wiki.lesswrong.com/wiki/Paperclip_maximizer http://wiki.lesswrong.com/wiki/Paperclip_maximizer A slight mis-prioritization on the AI's goals could easily lead to spectacularly unintended results.
- lubos 16y ago
  it's martingale betting but you have a point
  - khafra 16y ago
    Fixed, thanks. If it weren't for the frequent malapropisms people might mistake me for a well-read and thoughtful person.
- jamii 16y ago
  True. I would imagine that the way to use AI in the context of managing information overload on the battlefield is to have AIs prioritise information and propose options and have humans review the options and take decisions. By having the AI deal with incoming information and humans making the decisions you get the best of both worlds, at least with current technolody. Taking humans out of the loop entirely won't be a good idea for a long time yet.
  - anthuswilliams 16y ago
    Keeping humans in the loop doesn't do as much as we think, because humans have a natural inclination to trust computers. In a high-stress situation, such as a battlefield, a human will defer to a computer even if the computer reports information that, under normal circumstances, would be obviously false. In 1988, for example, the USS Vincennes shot down a civilian airliner because its AI targeting system had misidentified it as an enemy aircraft. Giving individual humans access to an off-switch is a pretty suboptimal way to prevent AIpocalypse.
jimbokun 16y ago
Human brains have highly evolved empathy software that can cause a human soldier to pause before taking an action that will end with a lot of casualties. True, the process of making soldiers involves a lot of de-sensitizing these instincts. But my point is those instincts are present in normal human beings and have to be over-ridden with extensive training, vs. robots that are likely to be intentionally designed to be psychopaths.
- philwelch 16y ago
  I'm not sure whether this is a point in favor of humans or in favor of robots. Inflicting enemy casualties is the point of warfare.
  - icegreentea 16y ago
    Winning is the point of warfare. Inflicting enemy causalities just happens to be part of winning. Ok, so rather than 'winning' it's achieving the desired change through application of force. If the point of warfare was inflicting enemy casualties, than we would have already won in Iraq and Afghanistan... and Vietnam. We've been moving away from the whole 'kill as much people as possible' method of warfare for quite a while. Omelets require breaking eggs, but the point of omelets isn't to break eggs.
jakevoytko 16y ago
I've worked with small robotic military platforms, and AI must overcome major physical limitations before anyone even worries about whether the AI is making good decisions! Battery life. If necessary, humans can stay active for a long time. Robots can't. If it dies, you either manually get the robot, or leave it for the enemy. The more processing and sensors on a platform, the sooner it dies. An almighty AI will likely require an almighty amount of power. Cameras. Humans swiftly adapt to lighting, can deal with very dark and very bright conditions, have an amazing field of view, can accurately track tiny objects very far away, and very quickly refocus and swivel our eyes and head. You'll have a lot of trouble finding cameras+mounts that meet all of these criteria that aren't a giant bullseye. It's more likely that you need to tailor your sensors for particular situations. What we see as a person at 80 meters might be a few pixels if the camera is zoomed out. If you zoom in, you'll see the enemy, but not your surroundings. Communications. Quickly transmitting binary data at the ground level is a very open-ended problem. Positioning. GPS doesn't work within buildings. Some of your platform power will be allocated to dealing with this fact.

by jamii 16y ago

khafra 16y ago

The stereotypical danger isn't that they'll make fewer mistakes than a human, it's that the mistakes they do make will be much bigger--continually offloading decisionmaking and decision enforcement to AI without a coherent theory of friendliness is like using Martingale betting: You don't lose often, but it only takes one to wipe you out.

jimbokun 16y ago
Much Paper Clip AI potential. http://wiki.lesswrong.com/wiki/Paperclip_maximizer http://wiki.lesswrong.com/wiki/Paperclip_maximizer A slight mis-prioritization on the AI's goals could easily lead to spectacularly unintended results.
lubos 16y ago
it's martingale betting but you have a point
- khafra 16y ago
  Fixed, thanks. If it weren't for the frequent malapropisms people might mistake me for a well-read and thoughtful person.
jamii 16y ago
True. I would imagine that the way to use AI in the context of managing information overload on the battlefield is to have AIs prioritise information and propose options and have humans review the options and take decisions. By having the AI deal with incoming information and humans making the decisions you get the best of both worlds, at least with current technolody. Taking humans out of the loop entirely won't be a good idea for a long time yet.
- anthuswilliams 16y ago
  Keeping humans in the loop doesn't do as much as we think, because humans have a natural inclination to trust computers. In a high-stress situation, such as a battlefield, a human will defer to a computer even if the computer reports information that, under normal circumstances, would be obviously false. In 1988, for example, the USS Vincennes shot down a civilian airliner because its AI targeting system had misidentified it as an enemy aircraft. Giving individual humans access to an off-switch is a pretty suboptimal way to prevent AIpocalypse.

jimbokun 16y ago

Human brains have highly evolved empathy software that can cause a human soldier to pause before taking an action that will end with a lot of casualties. True, the process of making soldiers involves a lot of de-sensitizing these instincts. But my point is those instincts are present in normal human beings and have to be over-ridden with extensive training, vs. robots that are likely to be intentionally designed to be psychopaths.

philwelch 16y ago
I'm not sure whether this is a point in favor of humans or in favor of robots. Inflicting enemy casualties is the point of warfare.
- icegreentea 16y ago
  Winning is the point of warfare. Inflicting enemy causalities just happens to be part of winning. Ok, so rather than 'winning' it's achieving the desired change through application of force. If the point of warfare was inflicting enemy casualties, than we would have already won in Iraq and Afghanistan... and Vietnam. We've been moving away from the whole 'kill as much people as possible' method of warfare for quite a while. Omelets require breaking eggs, but the point of omelets isn't to break eggs.

jakevoytko 16y ago

I've worked with small robotic military platforms, and AI must overcome major physical limitations before anyone even worries about whether the AI is making good decisions! Battery life. If necessary, humans can stay active for a long time. Robots can't. If it dies, you either manually get the robot, or leave it for the enemy. The more processing and sensors on a platform, the sooner it dies. An almighty AI will likely require an almighty amount of power. Cameras. Humans swiftly adapt to lighting, can deal with very dark and very bright conditions, have an amazing field of view, can accurately track tiny objects very far away, and very quickly refocus and swivel our eyes and head. You'll have a lot of trouble finding cameras+mounts that meet all of these criteria that aren't a giant bullseye. It's more likely that you need to tailor your sensors for particular situations. What we see as a person at 80 meters might be a few pixels if the camera is zoomed out. If you zoom in, you'll see the enemy, but not your surroundings. Communications. Quickly transmitting binary data at the ground level is a very open-ended problem. Positioning. GPS doesn't work within buildings. Some of your platform power will be allocated to dealing with this fact.