We expect to access chatGPT and DALLE and other generative AIs at scale. Like Twitter or a Google search, we don’t expect access to be limited. No matter how many people want to tweet or search at the same time, we expect the system to handle it, no waiting: instant result!
Right now, generative AIs’ scales are limited: chatGPT may be closed to you during peak usage hours, and Lensa took about half an hour to produce 100 terrible images that mostly don’t look like me (despite being based on twenty actual photos of me which I recklessly uploaded to the app).
Humans tend to assign value to the rare and not to the common. We call evolution’s most successful species, like rats and pigeons, “pests,” and regard evolution’s great failures and ball fumblers, like dinosaurs, dodos, and pandas, as marvelous treasures worthy of obsession. We considered the passenger pigeon a pest when they blackened the skies, then hunted them to extinction, and then spent a century talking about how terrible it was that we destroyed this once-great species that we used to regard as airborne filth.
It got me thinking about accounts I’ve read of Eliza, one of the earliest chatbots, which gets referenced in recent books by Sherry Turkle and Shoshana Zuboff’s (essential) Age of Surveillance Capitalism, and others: Eliza was programmed to give standard "psychologist" talk-therapy responses, and its users — apparently! — often said they felt like it was a caring therapist. But that was back in the 60s and 70s, when access to Eliza was exclusive.
The question I ask myself is whether today's humans want to experience empathy from a machine, since today it feels less special and more like a substitute without social sustenance — like a pacifier instead of milk.
I think of early CGI in movies, and how the then-innovative visual effects in The Matrix became common and accessible and therefore played for laughs rather than wows. The technology used today to create a Matrix-referencing punch line is better, quicker, and easier1 than what was used to create the original film, and so, socially, the tricks have lost their cachet.
There’s also the fact that when new effects are used, they dazzle us and seem so real — yet when we look at them many years later they fall flat. Have you watched Shrek lately? It was revolutionary when it was new — so realistic!! — now it’s kind of basic-looking.
Who Framed Roger Rabbit was really neat when it came out in 1988, but now, instead of thinking, “wow this looks so real” you’re more likely to chuckle at how fake it seems. Visual effects improved over time, and even if we couldn’t quite explain why the new effects were better, we nonetheless could better perceive the flaws in the old.
We may also grow less easily impressed (or deceived) by generative-AI’s early versions, as they become ever more convincing in their newer iterations. As we ride that knife’s edge of credibility moving forward, always finding the new thing more persuasive than the old thing, I suspect evolution’s energy-conserving innovation called “the human animal” will tire from frustration if we cannot determine who is a socially-relevant node in the human network and what is a digital placeholder algorithmically, generatively, pretending.
We don’t just want to know if the “voice” we’re communicating with is attached to a body, we need to: we need to know if we can marry it, or fight it, or compete against it in a race. We need to know if it could steal our lover’s heart from us. Can it can get sick, or die? Can it own or sell property? Can it hire or fire? Can it introduce us to a new world of friends? Will it come to our funeral?
We need to know these things to be social beings. If we can’t tell who’s socially relevant and who’s a ghost, laziness itself will drive us away from what were once our digital pleasures.
I've noticed in comics lately that many popular new works are drawn more roughly and people are especially committed to hand-written lettering, even where it's harder to read. I think that's partially a backlash against the very-polished-looking but zero-effort "works" of AIs. I think there's a desire for signs of human-ness in the made-thing. We'd rather see errors and unevenness and know a real person is talking to us than be dazzled by slick machine productions. And I think that's because we (simply!) want to live in a society of humans.
No matter how much you scale a technology, the human desire for social hierarchy is going to kick in. That’s why famous people have their own social apps, rich people have their own dating apps, journalists felt real feelings about their twitter blue checks, and one current form of youthful rebellion is to refuse technology altogether.
We are people, and people want to have status and set themselves apart — even if within the confines of a very narrow affinity group. The capabilities of AI mean that they will begin taking the places of fellow humans in the human network; they will be nodes in the human network that happen to not be human. What will our status be in comparison to them?
If status is measured as it is today measured online — by engagement, clicks, followers, and so on — AIs are exquisitely positioned to render us all no-status schmucks.
My experience is when humans are confronted with a game they can’t possibly win, a good number of them are going to flip the card table. (Which can also mean: Pitch the tea; or Summon the guillotine.)
Status is also measured in other ways: economics measures status with numbers next to dollar signs — and this is powerful. Online status became important enough to fight about when that status could be turned into cash.
AI can operate as an army of powerful agents, working to help someone bring in big bucks and achieve big status; but we could also find ourselves faced off in competition against those same agents; our status could be annihilated by them.
If you were an American autoworker in the 1980s, you’ve already danced to this song: you had a home and a family and property and pride. You were someone who mattered in your social network. Your father and father’s father worked at the plant and you had every reason to think your kids could get good jobs there too.
And then, another human benefited from an army of robotic arms, by setting them in competition against you.
The human who made that decision moved into a mansion with a granite foyer and a pool and more bathrooms than one family can use, its big windows looking out over the lake; you watched your neighborhood crumble. The robot arm stops only for scheduled maintenance.
“Which way will society go?” we might ask, but the answer even to how did society go will depend on your vantage point. More practically we might ask “Who will be society’s winners?” because we want to try to play the new game to win, or help our kids win.
But that’s misguided too: so long as there are “losers” at this scale — whole communities deprived of dignity they thought they owned — the whole society loses, even if some people end up with more digits in their bank accounts, because we all reap the plight of social instability together.
You can’t buy your way out of crime, unrest, divisiveness, and widescale social mistrust. We’re all connected. We cease to trust our corporations, and that comes back on us when they start laying off workers. We cease to trust elections, we’re left with conspiracy theorists for candidates. We boycott our religious institutions and we no longer have a voice there. We dismiss the value of education, and we end up with taboo and forbidden words that teachers must never speak.
If we’re smart, we’ll protect one another, because we are flesh-and-blood humans who need love and pride and air and water, who feel pain, who can die. If we are smart, we will cherish the imperfect human-ness in our art, music, and writing. The machines can mass produce the slickly glazed iconic things.
Someday, they may even imitate our individual quirkiness, a kind of “artisanal AI” that costs a little more, for those who can afford it.
But hopefully our eyes and ears will stay sharp, ahead of that knife’s edge of innovation, and we will grow to cherish ever more our human imperfection.
(Gone)
In 2019 the first year college students in my Conversation seminar concluded during a group conversation that what made something “meaningful” was effort, that the very essence of “meaning” in life was confronting difficulty. The more I’ve thought about this, over the years, the more I feel those students nailed an essential truth about life. If they were right, every effort-saving innovation runs the risk of directly depriving our lives of meaning.
Waiting for Artisanal A.I.
This is the most beautiful thing I’ve read yet about our little AI fixation, and so obviously written by a human.
Just came across your substack via the Convivial Society. I read Reclaiming Conversation and Alone Together when they were first published and was struck then that we needed to hold on firmly to what makes us human. In the homeschool co-ops I have been organizing for the past ten years, one of the rules was "no phones"; this included students and parents hanging out. Interestingly many of the parents found it really challenging, but they appreciated the mental break they got from being away from their phones. If parents don't model it, children can't mirror it. I just recently posted an article on 'TikTok brain cure with three ingredients' https://schooloftheunconformed.substack.com/p/tiktok-brain-cure-with-three-ingredients, focusing on the importance of real life relationships:
"...people who give ‘likes’ or contribute to viewer count are tokens, they are not people that we are in relationship with. Surround yourself with actual people. Relationships are sustained not because we uploaded the perfect picture or said just the ‘right thing’, but because we invest time, give part of ourselves, make sacrifices, demonstrate faithfulness, share in others pain, joy, struggles, hope, and care enough to do it again. At times virtual connections can act as a surrogate, but they are not the real thing. Strive to make relationships real, because relationships are at the core of what makes us human."