Is Generative AI Alive?

Source: Taylor Brandon/Unsplash

Cormac McCarthy closes his masterpiece, Blood Meridian—an anti-western that depicts in harrowing detail, and stunning prose, the horrors of the mid-19th century scalp trade—with an enigmatic epilogue describing an anonymous “man” making holes in the earth with a steel implement that “strik[es] the fire out of the rock which God has put there." Accompanied by a group of wanderers searching the prairie for bones, the technician and his associates “move haltingly in the light like mechanisms whose movements are monitored with escapement and pallet so that they appear restrained by a prudence or reflectiveness which has no inner reality.” They act, that is, like machines—“escapement and pallet” being the parts of a clock that lock and unlock the gear train in a precise rhythm, producing the clock’s melodic ticking—governed by a preordained “sequence and causality,” rather than an inner volition of their own.

This seems an odd way to end a book about greed and unimaginable violence on the frontier, until one remembers a reflection offered chapters earlier by a reclusive hermit living on the outskirts of society: “You can find meanness in the least of creatures, but when God made man the devil was at his elbow. A creature that can do anything. Make a machine. And a machine to make the machine. And evil that can run itself a thousand years, no need to tend it."

Perhaps, then, Blood Meridian ought not to be read as a commentary on the inner depravity of the human condition but on our capacity to create machines and systems that gain power over us by remaking us in their image and then take on a life of their own. Is the man at the end of the novel a Prometheus using his “implement with two handles” to steal God’s fire from the earth, or does he serve at the pleasure of his device, an instrument that compels him to carry forth the evil it was designed to inflict—an evil that violates the earth by breaking open “the rock which God has put there” and mining it for profit?

Source: Mike Surber / Center for Psychological Humanities and Ethics

A similar question is taken up by the contemporary philosopher John Manoussakis in the talks—soon to be collected in a single volume—he offered as part of the prestigious Fenwick Lectures back in 2019. Noting that “[i]f all life is generative, human life alone is also creative,” Manoussakis makes a distinction between procreation, which produces a copy of the individual (and thus the species to which it belongs), and creation, which produces something different or new. This insight—which was perhaps first offered by Plato who, in the Symposium, reflects on the difference between one’s children and one’s artistic creations—can be easily understood if we contrast the generation of human beings with the invention of tools and technologies. Our offspring, it is plain, resemble us. Our devices do not. And for Manoussakis, the key distinction resides in this: Generation gives rise to living beings whereas creation can only produce inanimate objects.

That at least would seem to be the case, though Manoussakis gives us reason to doubt his initial assessment when treating the work of the 20th-century computer scientist Norbert Wiener. Hailed as “the father of automation and cybernetics,” Wiener held that not only can our computers learn, change, and reproduce, they are as alive as you and me. Asserting that Wiener’s work introduces “a new, Darwin-like paradigm shift,” Manoussakis writes, “As Darwin had erased the demarcation line that sharply distinguished the human from the animal, so does Wiener’s essay with respect to the line that separates the human from the mechanical.”

Interestingly, Manoussakis’s lecture—delivered six years ago—and Wiener’s much earlier writings—resonate with questions arising from experiences we are only just now having today. In a recent article in The New Yorker, “Will the Humanities Survive Artificial Intelligence?,” Princeton historian D. Graham Burnett writes about his attempts to creatively integrate generative AI into his classroom by getting students to engage ChatGPT in a conversation on the history of attention. Treating the tool not as a computer but as a conversation partner, his students are surprised by the wisdom and depth of feeling it seems to evince. After reviewing their papers on the experiment, Burnett reflects:

Reading the results, on my living-room couch, turned out to be the most profound experience of my teaching career. I’m not sure how to describe it. In a basic way, I felt I was watching a new kind of creature being born, and also watching a generation come face to face with that birth: an encounter with something part sibling, part rival, part careless child-god, part mechanomorphic shadow—an alien familiar.

Prophesying the advent of this exact encounter, Wiener wrote back in 1964 that our machines may one day “develop an uncanny canniness.”

Such an alien familiarity is likely due to the fact that our devices not only learn to resemble us, but we learn to resemble them, conforming ourselves to their image and likeness, often in disturbing ways. Noting that machines are becoming increasingly capable of reproducing themselves, developing their own procreative capacities and thus giving “life” to new generations of machines (and also transforming we who emulate them), Manoussakis writes: “Life is of course a phenomenon notoriously difficult to define—but if machines reproduce, interact, adapt, respond, and communicate, what do we call them? And, more importantly, what do we call ourselves who created them?”

This is a difficult question, one made more complex in light of recent reports of users developing delusions and spiraling into ChatGPT-induced psychosis. It is, however, one that we must confront, especially if we hope to retain some semblance of “inner reality” and resist becoming mere cogs in the machine of an evil that, as McCarthy warned, “can run itself a thousand years.”