October 29, 2017

POSSIBLY IRRITATING ESSAY: What If Super Intelligence Just Means “Super Mistakes”?

Using the Programme Guide of the World Science Fiction Convention in Helsinki Finland in August 2017 (to which I will be unable to go (until I retire from education)), I will jump off, jump on, rail against, and shamelessly agree with the BRIEF DESCRIPTION given in the pdf copy of the Programme Guide. The link is provided below…

What would a superintelligence, artificial or natural, be like? Would it be able to solve the same problems we can but quicker, or would it look on the world in a completely different way? What would the consequences be if we shared the world with one, or more than one? Do we want to build something smarter than we are? But do we have to [in order] to survive?

Tom Crosshill: SF, F, and YA Contemporary author
Benjamin C. Kinney: Escape Pod Assistant Editor as well as neuroscientist and writer
Mikko Rauhala: short story author in Finland, MS in Computer Science/Intelligent Systems, transhumanist, singularitarian, Research Benefactor for the Machine Intelligence Research Institute, information society civil rights activist, member of the Electronic Frontier Foundation, vegan
Lettie Prell: Science fiction and fantasy writer

Why does everyone make the assumption that super smart equals Savior Of All Mankind?

That seems to be the current flow in the speculative fiction world. Gone are the explorations of the dark side of AI (Starting HAL in 2001: A Space Odyssey – a list of movies for your deliberation – https://www.wired.com/2009/01/top-10-evil-com/), hello, “I Sing the Body Singularity!” and features people consistently hoping that they’ll be around for the event.

In case you don’t know what the Singularity is, the brief definition: “the hypothesis that the invention of artificial superintelligence will abruptly trigger runaway technological growth, resulting in unfathomable changes to human civilization.” Jon von Neumann, Vernor Vinge, Ray Kurzweil, Eric Horvitz, and many other SF writers are typically proponents of superintelligence; major detractors include Stephen Hawking and Elon Musk as well as the works of Frank Herbert (the rebellion against AIs was called the Butlerian Jihad) and both incarnations of BATTLESTAR GALACTICA.

Even so, while the detractors assume malicious intent of the superintelligence, I wonder if it might end up less dramatic than that.

In the school I work at, I’ve known incredibly intelligent young adults who are poised to make their mark on society in a powerful way – one in computers, one in biophysics, another in medicine, one even in particle physics. I have also known incredibly intelligent young adults who were poised to make their mark on society and who failed spectacularly.

Why do both the proponents and detractors of AI believe in an intrinsic “good/bad” or good vs evil orientation. What if the Singularity produces both spectacular advances – and spectacular failures? Are we prepared for that?

I highly doubt it. We already have trouble following metaphysical gods. Most speculative fiction writers will instantaneously point out the number of murderers who thought they were hearing “God’s Voice” as he directed them to murder abortion rights advocates or even rock-and-roll singers. I have yet to read a story in which a speculative fiction admits to the entire history of both the old Soviet Union and present-day China (as well as several other countries) are officially atheist and summarily order the slaughter of Tibetan monks, underground Christians, Nazi Germany (while technically Lutheran…which we also never mention) ordered the extermination of Jews, jailing of devout Muslims...

What makes people believe that an artificial intelligence will be any better at “leading Humanity” than a metaphysical God? There will, of course, be “believers”. I had an odd thought while writing this that among the proponents of AI and superintelligence, there may be some who are doing the same thing many Americans do, particularly those in the 1950s like teen fiction SF writer Robert A. Heinlein who possibly went along with the prevailing religious mores of the dominant society in order to get published. What if individuals convinced of the coming superintelligence wrought apocalypse of all “normal Humans”, make certain that their “approval” of the Singularity is well-documented so that our new root overlords or superintelligent masters leave them alone? Perhaps the True Believers will be coopted by the superintelligence(s) to, you know, “run stuff” for them, the SI putting them in charge of the everyday operation of the world.

Nevertheless, the problem is that we constantly second-guess whoever is "over" us -- presidents, popes, generals, and head football coaches -- are criticized by the ones being led (or entertained). Once a superintelligence evolves, what's to keep all the scifi typed from bemoaning the job they're doing? What's to keep a superintelligence from making super mistakes?

Nothing.

Those cheerleading for superintelligence today might get exactly that. They may be pushing for it in fact. Of course, those operating at that level might be treading thin ice. Collaborators fare poorly in Human society and superintelligence won’t have anything BUT us to model itself/themselves after initially taking over the world.

So…superintelligence – something to cultivate or inhibit?

If we choose to inhibit it however, well, we all know what happens as soon as you tell people they “can’t do that”…


No comments: