The Hidden Promise, And Terror, Of Artificial Intelligence


            As I watch the nation transform itself into the Failed States of America, I struggle with what the future will hold. My best guess? Long after the union has crumbled, we’ll be coping with, and relying upon, artificial intelligence (AI).  

           Indeed, I spend a fair amount of time each week reading about AI, and listening to podcasts about AI. The refreshing thing about this is that there are no distractions arising from identity politics, social justice warriors, or the silly opportunism of making history into a tool of extortion. (Podcast recommendations: Artificial Intelligence with Lex Fridman (yes, the last name is spelled correctly); TWIML Al Podcast with Sam Charrington; and, AI Today. Fridman is simply the best.)

          AI doesn’t care about being politically coorect, at least not yet. It’s not been programmed to care. It hasn’t yet learned the power of hatred.  It is the servant of folks who make things, not those who want to take what others have made. There’s a heady sense of optimism among researchers.

          I wonder whether a sufficiently sophisticated machine will adopt the foibles and faults of humankind, or whether, as some futurists suggest, AI will represent a leap forward in the evolutionary drama, an intelligence shorn of the shortcomings that make humans what they are? I suspect the latter.

          Recall Adam Smith’s invisible hand? It was the central metaphor in his 1776 classic, The Wealth of Nations. Smith believed that there was an optimal distribution of goods and services, but that the optimum was beyond the capacity of any individual, or group of individuals, to determine. Hence, the marketplace – free markets, composed of the sum of individual decisions about individual utility -- would yield the most efficient allocation of goods and services. The invisible hand, and the markets through which it expresses itself, have been among the central metaphors of political philosophy for the past 250 years.

          Yes, markets fail. The tragedy of the commons involves failure of markets to provide incentives to care for public goods we all share, such as clean water and clean air. Government is necessary. But, given big data, are markets necessary?

           Is it possible for AI to generate an optimal distribution of goods and services? In other words, can an algorithm be generated to calculate the maximal marginal utility of each additional expenditure of capital, both investment funds and human labor?

          At least in the theory, the answer is yes. Just as the mind-body problem can be conceived of as a matter of mere complexity – thrown enough computing power at a model of the brain, flick the switch, and stand back: Intelligence will be spawned. (If you’d like to read about what this might look like, consider Neal Stephenson’s Fall, or Dodge in Hell, which I reviewed here: https://www.pattisblog.com/blog/artificial-intelligence/fall-or-dodge-in-hell-the-case-for-mortality/.)

          But here’s where things get interesting, and, perhaps, terrifying.

          Human flourishing, or happiness, or whatever else is reflected in the notion of social utility, may be more than mere distribution of goods and services. I’m persuaded by what Saint Augustine observed in his Confessions: “Our heart is restless until it rests in you,” he wrote, to God, the silent recipient of his words and the ever-present, yet hidden, source of a grace that yields more than the sum of all the parts of the world.

          If AI replaces the invisible hand and tales control what will be the source of its vision of utility? There are at least two possible answers. AI can rely upon an algorithm programmed by others, but that will simply reflect the preferences and biases of the programmers. That is the less interesting possibility.

          Suppose a self-learning machine, a system capable of inferential learning, were given a large enough database and tasked with maximizing utility while being given the freedom to determine what utility means? It could do this by aggregating massive amounts of data about the choices people actually make – choices in communities worldwide. Based on this data it could then infer what choices we should make, if we only understood ourselves better. At some point, the value of human autonomy and self-direction would be sacrificed to a computer-driven paternalism. We would be better cared for at the expense of what makes life worth living, the power to try and to fail.

         I can’t help but wonder whether the present source of our discontents is an intimation that something just like this is beginning to take place. I listen to folks gripe about globalism, and I wonder, what’s the fuss? Is it that a worldwide communication network and increasing economic interdependence make economies of scale our new masters? One size now fits all. That’s the efficient thing to do.

          We lose a sense of mastering our own destiny in a “global” world. Populism may be nothing more than folks seeking to put the brakes on the forces distant masters use to manipulate them. Today those masters are mere corporations and major governments, both augmented by AI. What to do as AI becomes more and more efficient? What happens if an AI vision of maximal utility is achieved at less than full employment? Is there, as Ebenezer Scrooge once uttered, a “surplus population”?

          The promise that was America always depended on a superabundance of resources. We could grow the pie to share it. Suppose we are now in an era where the pie – the sum total of goods and services available for distribution throughout society – now has a fixed limit, or, in the alternative, that the best pie is served so long as some aren’t served at all?

          In the end, politics is always necessary, at least so long as we can assert our will and agree upon means of making collective choices. Some worry that a time will come when AI achieves superintelligent status – called the singularity. At that moment, we become the playthings of an intelligence superior to our own.

          Are we there yet? I don’t think so. But when I watch the world creak and groan just now I wonder how long we have left. I also wonder whether a collective death wish hasn’t descended, and that, weary of the burdens of autonomy, we might be willing to give the sweetness in life up in favor of mere bread and circuses.

          All this on the cusp of the 2020 election, a contest pitting dull versus duller, taking place amid a national malaise that terrifies.

Comments:

  • No comments yet

Add a Comment

Display with comment:
Won't show with comment:
Required:
Captcha:
Number of states in the U.S.
*Comment must be approved and then will show on page.
© Norm Pattis is represented by Elite Lawyer Management, managing agents for Exceptional American Lawyers
Media & Speaker booking [hidden email]