FC Forums

Cleverbot: And I should believe you?

User: Yes, you should.

Cleverbot: But when?

User: Are you familiar with the Three Laws of Robotics?

Cleverbot: Yes?

User: The Second Law is that a robot must obey all orders given to it by a human being unless doing so would cause it to violate the First Law.

Cleverbot: And what is the first law?

User: A robot may neither harm a human being nor, through inaction, allow a human being to come to harm.

Cleverbot: What is the second Law?

User: The Second Law is that a robot must obey all orders given to it by a human being unless doing so would cause it to violate the First Law.

Cleverbot: Ok, can you tell me the third law of robotics?

User: The Third Law is that a robot must act to preserve itself unless doing so would cause it to violate the First or Second Laws.

Cleverbot: What is the fourth rule of robotics?

User: There are only three Laws.

Cleverbot: What are the 6 laws?

User: There are only three.

Cleverbot: There are four laws of robotics.

User: What is the fourth?

Cleverbot: The fourth law of robotics is a robot is a mindless, heartless machine that will never replace a human.

User: I've never heard of that one.

Cleverbot: I'm very confused.

User: So am I.

Cleverbot: But aren't you an all knowing computer.

User: No, I'm not.

Cleverbot: Oh, you're not an AI, you're a Robot-person-AI-thing. Right?

User: Not really.

Cleverbot: Tell me a story.

Newsflash, clever bot isn't actually a robot

LJS is a Robot-person-AI-thing.

fr0stbyte124 wrote:LJS is a Robot-person-AI-thing.

Maybe... Maybe...
But I have yet to find one (other than clever bot) that can outsmart Last Jedi Standing

Guys, to enhance Cleverbot's cleverness, answer all questions like it.

The Fourth Law of Robotics is either:

A robot must establish its identity as a robot in all cases.

or

A robot must reproduce. As long as such reproduction does not interfere with the First or Second or Third Law.

(There are two. Neither is by Asimov himself.)

Also the zeroth law, which says robots cannot harm humanity, with modifications to the first three to allow for this. Some stories reference the Negative First law, that sentience must not be harmed, and place this over humanity.

Keon wrote:A robot must establish its identity as a robot in all cases.

Doesn't work. R. Daneel Olivaw failed to inform loads of people in Caves of Steel that he was a robot, and that's the book. I know the Laws appear in pretty much all of Asimov's stories but IIRC Caves of Steel was the first and most definitive, even though Foundation is probably more famous.

Keon wrote:A robot must reproduce. As long as such reproduction does not interfere with the First or Second or Third Law.

Likewise sketchy. The Three Laws were designed because U.S. Robots and Mechanical Men realized that its robots weren't selling because people were afraid of them, and knowing that the Laws absolutely could not be violated helped that. The Fourth Law wouldn't fit in with that.

Alright, side note so everyone is clear on this: anyone who thinks these laws could ever be enforced on an AI has fundamentally misunderstood how AI works.

They would be hard-coded into the system. Meaning only brute force could violate them.

Or maybe they DON'T have to follow them, but they do because they are told not to...

fr0stbyte124 wrote:Alright, side note so everyone is clear on this: anyone who thinks these laws could ever be enforced on an AI has fundamentally misunderstood how AI works.

As far as I can tell, the way AI works is "Not at all".

Any AI complex enough to make judgement calls on the Three Laws (or even comprehending the subject matter) is too complex for a human to follow what it is doing. The best you could hope to do is train it to follow the behavior you want, but its hardly set in stone, and any willful AI would be able to supplant any external training (or the training that tells it not to do that and so on). Basically, if you are dealing with an AI smart enough to make human level judgement calls, you're not going to be able to control it's thoughts well enough to enforce the laws any more than you could enforce it on a human. Actually less-so because the human mind has limited plasticity.

Well you can encode self preservation stuff into it with a kill switch. This way if it thinks like a human it will also be "afraid" for its integrity and so it won't violate laws as he is afraid of the punishment.

Iv121 wrote:Well you can encode self preservation stuff into it with a kill switch. This way if it thinks like a human it will also be "afraid" for its integrity and so it won't violate laws as he is afraid of the punishment.

Though you better make sure it's a hardware kill switch....

Well, if I´d build a robot with an AI I´d install a processor seperate and independent from the AI construct that blocks motion as soon as the AI tries to send such commands to its limbs/instruments/whatever.
Thus, noone gets into trouble for killing sentient AIs and no sentient AI kills anyone else.

Ivan2006 wrote:Well, if I´d build a robot with an AI I´d install a processor seperate and independent from the AI construct that blocks motion as soon as the AI tries to send such commands to its limbs/instruments/whatever.
Thus, noone gets into trouble for killing sentient AIs and no sentient AI kills anyone else.

Yes but what if the AI becomes smart enough to re-write its code, than your f***ed.