Morality and Machines

Morality and Machines
Who teaches the machines? What is moral intent? When is it "good"?
David Amerland brings up questions about morality in the fragmented landscape of robotics. 

A thought provoking read with a lively and provocative comment thread. 

#morality   #machinelearning  

Originally shared by David Amerland

Morality

For anyone who has ever said “guns don’t kill people, people kill people” as an end to the argument about gun regulation the very suggestion that there is an aspect of technology that has an element of morality to it, must sound like anathema. Considering that the argument about guns, as employed, is itself fallacious: https://goo.gl/QaXThI it’s worth exploring whether technology itself should be subject to moral clauses. 

The use of technology is, it seems, now inseparable from the studies that show just how the brain is affected by it: http://goo.gl/wiaGGz. Where the brain is affected morality may not be far behind, especially where technologies like Facebook are concerned: http://goo.gl/jnUhxc, though the questions (and perhaps, charges attending them) can be levelled at any social media network. 

To show the complexity involved when examining this question consider the fact that a discussion on Reddit, on the subject (https://goo.gl/XWw0e2) started talking about machine morality and machines, evolving morality (!). 

An attempt to define purpose (and divine intent) with technology suggests that it cannot be morally neutral: https://goo.gl/H6vzlG. At the other end of the scale stand those who believe that ultimately it is people who make the difference, not technology: http://goo.gl/HQ4i7M. 

But as machines become more autonomous and begin to use information to make choices the ability to apply a moral code of some kind is beginning to feature in the discussions taking place: http://goo.gl/Oez8EY. A New Yorker article on “Moral Machines” talked about the possibility of intelligent devices that will have “ethical subroutines” - http://goo.gl/1BaOfv. 

Immanuel Kant considered that morality is defined as well as dictated by a categorical imperative that makes it impossible for us not to act in a specific way as humans: https://goo.gl/564ofQ.  Kant’s call to reason to buttress his argument also lends itself to it being applied by machine logic with some very interesting results: https://goo.gl/kmaV3V which suggest that when it comes to reason, reasoning, morality and machines we’re just not out of the woods, just yet. The reason why is given by Anthony Beavers who, in his treatise, considers not just morality but also ethics and how the two are actually closely connected and cannot/should not be divorced: http://goo.gl/x6HQ5i. 

The question of morality and ethics raises the thought of whether machines can truly be autonomous moral agents (AMAs) and what that would then mean for humanity: http://goo.gl/cYeJnj. As we are heading towards a robot society, the question of ethics and moral agents suddenly becomes more than just playful speculation: http://goo.gl/3eBcYO. 

Some, like George Lucas, seem to think that a machine can exhibit morality without having a conscience: http://goo.gl/CdmtPU while an essay at The New York Times suggested that the question of morality and machine code is now part of everyone’s thinking even if not many people believe machines can be moral: http://goo.gl/KwXSl. 

In his TED Talk, Nick Bostrom, suggests that we should start thinking about machine morality right now: https://goo.gl/6CwGy3 and, by association, we need to define a little more rigorously our own sense of what’s moral. To muddy things a little more consider the argument of whether morality is absolute or subjective: http://goo.gl/I4CZ5o and whether it is relative but not subjective: http://goo.gl/RELWos. 

Those who see the world through a theology obviously suggest that morality is an absolute standard that somehow we discover or manages to surface in us. Bioethicists, on the other hand, are looking at morality as part of a complex, evolutionary process that has resulted in our being here: https://goo.gl/OzEcDd.

Because morality is not really an integral part of intelligence: http://goo.gl/r80E2 the question of values and how they are created does come up. Alva Noe, who studies philosophy at Berkley (https://goo.gl/kIunkT) suggests that the moment we start talking about machine shaving “our values” as Nick Bostrom did, we are really talking about making safe appliances. Intelligence, he suggests, will have its own values: http://goo.gl/x1a2Xq and whether those coincide with ours will be the subject of conjecture. 

Our technological revolution is advancing at a pace that each of us feels on an almost daily basis. The fundamental arguments and debates around it however have barely started. The only thing that will determine the kind of future that is being created by our efforts will be our ability to tackle its challenges with an open mind and a largely consensual, networked, race and creed-agnostic approach. If we manage that, maybe … just maybe, we will also succeed in creating autonomous moral agents that will not be sociopathic killers hell-bent on exterminating the human race. 

While we are still here, of course, and still enjoying our role as Apex Predators, we need to make sure that coffee aplenty, croissants, cookies, donuts and chocolate cake are there to be had. Have an awesome Sunday, wherever you are.