Artificial Intelligence – Could it be Subject to “Garbage In Garbage Out”?

Big announcements around ChatBot GPI, Bing, and Google have sparked what one columnist calls “gotcha games” to fool the bot into doing stupid things.  One induced a bot to assume a female personage which tried to persuade him to cheat on his wife.  Needless to say, there are bigger issues around artificial intelligence.

A large overhanging question is whether AI programs can actually become conscious agents.  What would this mean for people, the only such beings we know of now?  Potential problems could include whether the new agents try to eradicate the competition (us); or whether human nature is anything more than the processing capacity endowed in these bots.

One feature of the AI bots now is that they still learn from the stimuli they receive.  This presumably explains the “female” that tried to seduce the journalist.  As bots become exposed to the public discourse, what will they learn, and how would that content shape the nature of their development?

What is out there for bots to learn from, and in what proportions among the various possibilities?  Note Holman Jenkins’ observation how the “gotcha function … long ago turned every politician into a scripted automaton.”  How many political ads are anything but fully predictable?  Or for that matter, ads for cars, liquor, or medicines?  If bots need repetition to learn their lines, or if they otherwise mimic human learning processes, shouldn’t we examine what dominates our sensory environment now?

If too much media, mass or social, is corrupting our minds, particularly of our young people, what would we expect of learning programs constructed by our own mental processes?  How much are those affected by ads, traffic jams, and pop music?  How much of our news comes in snippets already pre-digested by formulaic practices of editors and publicists?  

America implicitly holds an image of human agency, which makes personal rights “unalienable” and relegates government to the role of securing those rights. This image differs from an old view, often called “Realist,” that casts humans as one more animal that exists to survive and propagate, with all our emotions and rationales serving as tools to that overriding end.  

We can well imagine AI bots replicating the human person depicted in the Realist view.  If we believe that view best explains our learning functions, re-creating humans is a fairly direct exercise, without the wild cards of our irrational (spontaneous?) impulses.  These bots could, as well, be more efficient than live humans at self-protection and propagation.  To the extent that humanity has some “something more” to its volition, and to the extent that its emotions and impulses and brain glitches are distractions rather than sources of innovation, the bots could be in a position to outcompete the people.  If it came down to a war of existence, Realism-infused bots could outcompete humans and make the Realist view true.  There lies a danger to note.

If somehow this elusive whimsy of the human spirit makes itself attractive to the AI learning machine, how might that affect the shape they grow into?  There is room for literally infinite speculation.  Some possibilities could be quite comedic, like the seductress.  Some could be supremely irritating, combining the stylized conversation of news anchors, restaurant servers and press spokespersons with formulaic tunes of boy bands and jingle writers.  But who would guide the bot into “good” choices of what to learn from – or what innate drive might the bot have, to know “good feel” from bad?

Ultimately, there is great room for “garbage in garbage out” to shape AI thinking, for better or worse.  Some injunctions come to mind.  First, to pay attention to the public atmosphere we create now, in our conduct and our choices.  If AI looks at all our foibles as natural intelligence, better effort to shape the atmosphere as we really want is direly needed.  Second, as these bots take on more functions around us, and they will, we should pay attention to what they do, how we guide them, and what tasks we need to keep for ourselves.  Maybe then we won’t fear replacement or domination by computer programs.  Maybe in this self-definition we will raise the level of self-respect in the world.  Maybe we will reat each other as whole and free persons.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: