[manifestsilence]

And you know how the 3 laws turned out in Asimov's world...

Seriously though, I don't think Asimov was being naive with his choice of laws and then showing how they were then subverted. I'm sure he was aware of things like Godel's proofs and the difficulties involved with fuzzy logic and AI.

Asimov kept his laws simple so they would work in a story, but was likely well-aware that real laws could be made more nuanced and that they inevitably would still fail in the ways he described.

Implementation is going to be very, very hard to get right, if at all possible.

[shadowgovt]

> And you know how the 3 laws turned out in Asimov's world...

Pretty well? If you're referring to the original book (and not the excellent Will Smith movie that deviated from the source material significantly but was extremely entertaining in its own right), the robots correctly surmised that humans lacked the wetware capacity for global-scale planning and took on that burden. The world at the end of Asimov's story is many things, but it isn't dying from an utterly avoidable climate change disaster because of a transcontinental tragedy of the commons.

(It is, of course, a fiction. But if we can't take away from the fiction "We should trust robots with global resource planning" without some critical thinking, we shouldn't take away "robots can never be trusted and will always betray humanity" without some critical thinking either).

[TheCoelacanth]

I would assume they're referring to the original short stories published in I, Robot, which has numerous robots gone awry stories, but not the the taking over planning for all of human civilization plot that you are referring to.

That plot was from books he wrote 40 years later to unify the Foundation and Robots series. Many fans consider it a rage-inducing retcon.

[shadowgovt]

I'm referring specifically to "The Evitable Conflict" (https://en.m.wikipedia.org/wiki/The_Evitable_Conflict), which I believe appeared in the original 1950 collection. It specifically includes the plot of the world coordinating machines causing some harm to come to some individuals to optimize the survival of the machines and the outcome for humanity as a whole.

I'll have to look up the other stories you referenced though; I'm unfamiliar with them and they sound fascinating.

[salawat]

>It is, of course, a fiction. But if we can't take away from the fiction "We should trust robots with global resource planning" without some critical thinking, we shouldn't take away "robots can never be trusted and will always betray humanity" without some critical thinking either.

Ironically, it isn't so much leaving the robots to their own devices that is the issue. It's what we'll talk ourselves into doing with them that concerns me.

Then again, I'm a Protomen fan, so I may be prejudiced in that regard.

https://protomen.com/

[larnmar]

Realistically the three laws are silly here.

Better laws in this context would be something like “police robots cannot be armed”. Then maybe some other restrictions about circumstances under which they can be used — eg do we want a police robot sitting on every street corner monitoring? (Answer: maybe?)