You Can’t Hack Experience
It’s been an eventful two weeks. Our President-in-waiting decided that extra-terrestrials are demons, our Secretary of War declared that we’re fighting a Holy War for Jesus (who may or may not be Donald Trump), and the Head of FEMA’s Office of Response and Recovery teleported fifty miles to a Waffle House in Georgia. Meanwhile, the Strait of Hormuz remains closed, thousands of Marines are on site in Iran, we’ve forgotten about Cuba (which we were starving a second ago), and the price of fertilizer and oil is skyrocketing just in time for the planting season and summer vacation. Not to be outdone, our disintegrating Commander-in-Chief managed to explain that America can’t be bothered with day care (or scams like Medicaid or Medicare), even as the cost of his golf games to the American taxpayer is topping a hundred million dollars.
I’ve missed a few things. A lot of things. As I said, it’s been an eventful two weeks. Still, looking over the landscape of our folly, it occurs to me that if we manage to make it through this political hurricane we’ve conjured up for ourselves, the next order of business may be to revisit Calvin Coolidge’s famous dictum, “the chief business of the American people is business,” because it hasn’t been working out so well for us. Because business is becoming our only business, and things are getting lost in the process. Big things. I’m talking about shared values as old as our species. And assumptions about reality. And truth. And what it means to be human, or even alive. Or not.
It’s a subject that deserves a short book or a thirty-page manifesto, not a three-page post, but I’ll take a swing anyway.
Recently I was stopped short by two pieces in The New York Times, which used to call itself “The Newspaper of Record.”
The first presented two stories, one written by a human being, the other by A.I., and asked readers to judge which was better. The second, a well-meaning opinion piece by novelist Andrea Bartz, worried that the publishing industry was unprepared for the upcoming erosion of trust between readers and writers due to the ability of large language models to mimic actual writers - like me, for example.
What impressed me about both (though the Bartz essay at least attempted to grapple with some of the issues), was how entirely they’d bought into the notion that because A.I.-generated facsimiles will soon be indistinguishable from the original, they’ll be comparable.
They won’t. Because they’re not.
We can skip the ontological niceties here; Kant and Co. can stay on the sidelines.
A poem or a play or a piece of music written by a human being - whether good, bad or indifferent - is the product of a sensibility shaped by its own, particular experience of living on this earth. That experience, that history of desires and regrets, of things heard, read, felt, seen, dreamed and forgotten and dreamed again, forms the soil, the substrate, from which a story or piece of music grows. An A.I.-generated story or song is an entirely different thing because it emerges from no experience. Because it has no history. Because it has felt nothing. Because it’s not human.
Deep language models can fake human experience by combining and mimicking the expressions of billions of people who have actually lived; they can assume a history, a voice, a soul; they can imitate life in all its particularity and convince us that they’re us, but until they’ve lived, they’re not. The A.I. story in the Times is a forgery, an act of theft, and so far, at least, the courts seem to agree, which is why Anthropic, a large language model that recently stole the works of thousands of writers (my own among them), is going to be coughing up a lot of money this summer.
Which brings me to Bartz’s concern that the publishing world (along with the rest of us), is unprepared for the deep fakes coming its way. I have no doubt this is true. Lazy or unethical writers like James Frey, whose fraudulent memoir was exposed by Oprah years ago, and who is quoted in Bartz’s piece, will use it to spit out work posing as their own because – guess what? – writing is hard work.
“I have asked the A.I. to mimic my writing style,” Frey says (though maybe he asked A.I. to say it for him), “so you, the reader, will not be able to tell what was written by me and what was generated by the A.I.”
All of which points us to a very important question: For whose benefit is all this?
Not ours, clearly. Sure, a handful of Freys may manage to sell A.I. generated books as their own before others jump on the bandwagon and readers (and publishers) get wise, but beyond their sell-out, the larger impoverishment is incalculable. Because we’re not just talking about writers and books and publishing – we’re talking about everything that A.I. seeks to replace: our concentration, our imagination, our ability to create something of our own and to learn from the failures that the work of creation inevitably entails.
The history of technological development has been the history of making our lives easier, and to an extraordinary extent, our inventions have succeeded in dulling the hard edges of the world. We’ve prolonged our life spans, limited suffering (for some, at least), made advances against disease. We’ve put gravity and distance on notice. So far, so good. I don’t regret the development of penicillin or the invention of the crosscut saw.
But the time has surely come to ask if our ever-greater ability to dull the world’s edges hasn’t begun to dull us instead. We seem to have passed a historical pivot point beyond which more is less. A sprinkle of salt on your food will enhance its taste; a cup of salt will kill you. Our ability to farm out our human strengths – our memory, our creativity – to machines, is becoming toxic. What was intended for our benefit now seems increasingly for the benefit of our inventions.
So . . . again: Who profits from this? Who profits from us outsourcing our jobs, our selves, perhaps even – as the growing number of cases of A.I. related psychosis indicates - our very sanity?
The answer is obvious: The relative handful of individuals (and their investors), who stand to make a great deal of money. Who will gladly colonize us all to make their next twenty billion, to be part of the ongoing up-suck which has drawn 70 trillion dollars from the American economy into the accounts of the top 1% over the last 50 years.
Their problem, of course, is that you can’t hack experience. You can’t replace the benefits of effort (and failure). Of human interaction. Of our reliance on the evidence of our senses.
All you can do is imitate them while impoverishing the original.
But that won’t keep them from trying. Or from hiding behind that now nearly meaningless word, ‘progress,’ or its slippery synonym, ‘advancement.’
Progress for whom? I want to ask. At what price? And does progress mean improvement, or merely the next thing?
In 1950, the world was on the cusp of a new technology that would change the world. It was called the hydrogen bomb.
Progress? For whom?