School Survival Forums
Robot Ethics - Printable Version

+- School Survival Forums (http://forums.school-survival.net)
+-- Forum: The Lounge (/forumdisplay.php?fid=34)
+--- Forum: General Talk (/forumdisplay.php?fid=18)
+---- Forum: Philosophy, Politics & History (/forumdisplay.php?fid=10)
+---- Thread: Robot Ethics (/showthread.php?tid=8421)



Robot Ethics - Bob Dole - 08-01-2009 04:05 PM

So, I just finished reading I, Robot by Isaac Asimov. It was a very interesting book. The basic premise of the book is that intelligent robots have been created by humanity. These robots are subject to the three laws of robotics.

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2. A robot must obey any orders given to it by human beings, except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

http://en.wikipedia.org/wiki/Three_Laws_Of_Robotics
A bit here for more reading.

Asimov then goes through the entire book tearing apart his own three laws. It's a rather strange thing to do with laws of your own invention, don't you think?

Anyways, here's something for you to ponder. How does one strengthen those laws so that such bugs do not occur? I'll add my own thoughts in a while.


Re: Robot Ethics - psychopath - 08-01-2009 04:20 PM

Even if these laws do work, I don't think they will absolutely prevent robots going haywire.

Why do humans want to survive and reproduce? It's because if we didn't, we wouldn't be around. As the internet takes over the globe and there's a bunch of programs spreading all over, we'll probably get some that can modify themselves(maybe randomly like mutations, maybe for another reason...) while self-replicating.

Kind of like the beginning of life. From a soup of code, there could emerge self-replicating and "selfish" programs that exist for no reason other than the fact that only the ones that replicate themselves pass on their code.

Eventually, these "viruses" may get more and more complex, or, if they never modify themselves, more complex ones may pop up. Who knows, maybe they'll even compete for resources(bandwidth for example).

Even later, we could get very intelligent viruses that are very committed to survival. These programs will not care about anything but the replication of themselves. If humans try to destroy them, they would counter it.

Maybe if we try to turn off the machines that they inhabit, they'll infect other machines to physically stop us from doing so, because these AI care only about themselves.

No Asimov law could stop such an entity.

There's also the possibility that nutcases would build machines that want to survive and do not follow such laws. This would mean skipping the years of "cyber evolution" that the viruses had to go through.

Also, as software gets more and more complex, we understand less of it. There could be unexpected results due to nobody completely understanding how a robotic system works.

However there's still ages till robots reach our level of intelligence. Quantum computing is in its infancy and the best robots out there have insect levels of intelligence.


Re: Robot Ethics - Liquid - 08-26-2009 06:19 AM

(Hi, I'm back Smile )

psychopath Wrote:As the internet takes over the globe and there's a bunch of programs spreading all over, we'll probably get some that can modify themselves(maybe randomly like mutations, maybe for another reason...) while self-replicating.

That's beginning to happen right now. There have been "Polymorphic" viruses for awhile now but there aren't so many because there are hard to write, "Metamorphic" viruses are the next step up the chain.

With polymorphic code, there is the important "core" code, and fluff non-important code along with it. When the polymorphic virus replicates, it randomly changes the non-important code, without changing the core of the virus. "Metamorphic" viruses randomly regenerate the core of the code, but still keep the algorithms running the same way, just using different code. With metamorphic viruses the non-critical parts of the code are usually still there but aren't "fluff" instead it might be code for other viruses, toolkits, data on how the attacks are going, IPs of servers to send info to, and other things like that- things that aren't essential but change on replication.