3 ms·
That doesn't follow it all. Sentience concerns an internal state of mind, and we simply do not know if simulating the behaviour of sentient organisms requires u
by Barrin92 6y ago
That doesn't follow it all. Sentience concerns an internal state of mind, and we simply do not know if simulating the behaviour of sentient organisms requires underlying sentience. It even seems unintuitive because me not being able to distinguish Stockfish from Magnus Carlsen doesn't imply that Stockfish experiences chess the same way he does.
A more accurate conclusion is that, if I can't tell the difference precaution dictates that I treat the robot as if it were sentient, just in case it is, but it by no means follows from observation alone. Just because my cat thinks the laser pointer is alive, doesn't mean it is. And like cats, humans are likely to be way too easily fooled by fairly simple systems that just emulate external behaviour. Story comes to mind of soldiers developing trauma after seeing bomb defusing robot spiders flail around.
"Every time it found a mine, blew it up and lost a limb, it picked itself up and readjusted to move forward on its remaining legs, continuing to clear a path through the minefield. Finally it was down to one leg. Still, it pulled itself forward. Tilden was ecstatic. The machine was working splendidly. The human in command of the exercise, however -- an Army colonel -- blew a fuse. The colonel ordered the test stopped. Why? asked Tilden. What's wrong? The colonel just could not stand the pathos of watching the burned, scarred and crippled machine drag itself forward on its last leg. This test, he charged, was inhumane."
https://www.washingtonpost.com/wp-dyn/content/article/2007/05/05/AR2007050501009.html https://www.washingtonpost.com/wp-dyn/content/article/2007/0...
- p1esk 6y agosimulating the behaviour of sentient organisms A human child simulates behaviour of people around her, gradually becoming more sentient. Is this fundamentally different from a neural network being trained on real world data?
- Barrin92 6y agoyes, because a human child is embodied in the physical world, has sense organs, runs on electrochemical processes, and as far as its mind concerned has (developing) structures that facilitate self-awareness. It's perfectly reasonable to assume you can make a neural net, or even some other form of 'AI' that lacks most of those structures but outwardly performs the same tasks, just like a calculator does maths without (most people would agree) being conscious about maths. solving problems and being sentient are not the same thing, is the basic point. Emulating behaviour is effectivelly intelligence, sentience is awareness.
- p1esk 6y agoOk, so let's put our neural network inside a robot with sensors, so it can see, hear, smell, taste and touch its environment. Let the robot wonder around and learn how to do various tasks (imitation learning, self-supervised learning, reinforcement learning, etc). Let's say after a few years of such training it starts to behave like it was sentient. How do you know what structures have formed in its multi-trillion parameter space? How do you know if those structures are fundamentally different from human brain structures responsible for sentient behavior?