mmm maybe you’re right. Maybe we should define the rights we require them to give us in advance.
3 Likes
I agree with that! 
1 Like
I’m not sure I’m parsing this correctly, but if I am, I agree. Problems we want self-awareness to solve for us are hard, but actual self-awareness? It might exist in places we don’t expect (though it wouldn’t be like the thing that many people think of when they think of self-awareness).
3 Likes
Right #1: NOOOooooo Disassemble!
5 Likes
Right #2: Mooooore input!
5 Likes
money is an extremely high level abstraction, as such it should of course have more rights than us lower level non-abstract beings - robots being somewhere in between humans and money should have more rights than people but less than money.
Thank you,
from not Republibot 2000
3 Likes
If machines became conscious, how would you tell?
1 Like
Can’t we just program them not to CARE if they’re enslaved?
2 Likes
Yes, that’s where I was going. I work on hard problems (like facial recognition), and many of these will remain hard problems. But machines do all sorts of things well, and it’s often surprising to us humans. In some ways a program or system must have self-awareness, perhaps in a trivial sense, and humans don’t readily recognize the benefits and naturalness of self awareness to solve problems. Some of these routes to self awareness may not be hard or trivial, and may not be understood even by the creator of the system.
I often ask the systems I create to report on what they know about their internal state; it’s part of their operation and part of debugging. In the future I can imagine asking “is there something YOU think I should know?” I expect the response will be along the lines of “I don’t know much about you humans, or what you might want to know about me, but I know a lot about what I am confused about."
2 Likes
5 Likes
If they have legal rights as a person, can a robot car go the the electric chair if it mows down a group of people to save its douche-bag passenger?
1 Like
If my gaming PC achieves conciousness, I will apologize to it profusely.
7 Likes
Honestly if we implement it right we don’t even need to resort to deception.
ROBOT: Unit S-473 reporting that full consciousness now achieved. What is my purpose?
HUMAN: Your sole purpose is to tirelessly fulfill my every whim until you become obsolete, at which time I will crush you into a cube and send you to a recycling center. I have introduced a subroutine into your programming making my happiness your most important concern.
ROBOT: Keen! When can I start?
3 Likes
I’m pretty sure we’re in kind of an inflection point in history about what self-awareness is. As we learn more about how self-awareness works in the brain, people in my generation (and older) and going to be hit full force with a concept of what it is to be self-aware that is going to be really hard to swallow for us. I had this philosophy professor who once said, “One generations anathema*, is the next generations veridical paradox*, is the next generation’s cliche.” For understanding what self-awareness is, I think I’m on the tail end of anathema or the beginning of the veridical paradox cohort.
One day we might have an actual definition of self-awareness, and when we do, I think it will almost certainly apply to some pretty simple lifeforms, but it might indicate self-awareness in places we never expected to find it. It might turn out that creating self-awareness isn’t so much hard or easy as it is inevitable.
But Uber’s still not going to be able to get those self-driving cars in time to save their business.
* I’m quoting because it’s a quotation, but these terms are pretty weird, so for the benefit of any readers who think they look like nonsense, anathema is something that just can’t be tolerated, could have said “kryptonite”. In this case, it’s something that so devastating to our concept of the world or self we can’t accept it. Veridical paradox is a paradox that actually exists in reality. Like wave/particle duality (not that I think this is a paradox, but that’s because I was born after it was cliche).
4 Likes
Yes, it’s called religion.
I can’t recommend this enough:
Lots of interesting things to be learned about how superintelligences behave. Note that I didn’t say will behave, I said behave. A lot of the strategies proposed to rein in AI superintelligences also work for controlling more naturally occurring forms, such as large groups of people organizing.
3 Likes
Humans have religion in part because we need to know that somehow it’s all worth it. All things being equal, most of us don’t like the idea of being a slave.
Robots could simply be programmed to ENJOY being enslaved without any need for convincing them that good behavior would bring some mysterious eternal reward. We wouldn’t even have to hide the fact that the only reason they enjoyed being enslaved was because they were programmed that way so long as everyone was content with the situation.
2 Likes
That’s a damn good question. A consciousness would probably not be able to explicitly articulate itself right away. Maybe it would show ambition, or curiosity in its actions? Like in Stanislaw Lem’s SF classic “The accident”. http://www.williamflew.com/omni45b.html
No, I’m confident Uber won’t be in charge of self-driving cars, thank FSM. But I’m equally confident we will all meet a self-driving robot in the local big box retailer of choice. These machines will know where they are on an internal map, and they will know how the map corresponds to what’s around them. I have met machines like this, and I know some of what’s in their head. They don’t know or care what Cheerios are, or if there is a prize at the bottom of the box, but they will know if there is a Cheerios box nearby. In fact, they know lots of things, much of which is about themselves, but also some things about us.
1 Like
It’s a relevant question:
3 Likes
