AI has a lot of problems. It helps itself to the work of others, regurgitating what it absorbs in a game of multidimensional Mad Libs and omitting all attribution, resulting in widespread outrage and litigation. When it draws pictures, it makes the CEOs white, puts people in awkward ethnic outfits, and has a tendency to imagine women as elfish, with light-colored eyes. Its architects sometimes seem to be part of a death cult that semi-worships a Cthulu-like future AI god, and they focus great energies on supplicating to this immense imaginary demon (thrilling! terrifying!) instead of integrating with the culture at hand (boring, and you get yelled at). Even the more thoughtful AI geniuses seem OK with the idea that an artificial general intelligence is right around the corner, despite 75 years of failed precedent—the purest form of getting high on your own supply.
So I should reject this whole crop of image-generating, chatting, large-language-model-based code-writing infinite typing monkeys. But, dammit, I can’t. I love them too much. I am drawn back over and over, for hours, to learn and interact with them. I have them make me lists, draw me pictures, summarize things, read for me. Where I work, we’ve built them into our code. I’m in the bag. Not my first hypocrisy rodeo.
There’s a truism that helps me whenever the new big tech thing has every brain melting: I repeat to myself, “It’s just software.” Word processing was going to make it too easy to write novels, Photoshop looked like it would let us erase history, Bitcoin was going to replace money, and now AI is going to ruin society, but … it’s just software. And not even that much software: Lots of AI models could fit on a thumb drive with enough room left over for the entire run of Game of Thrones (or Microsoft Office). They’re interdimensional ZIP files, glitchy JPEGs, but for all of human knowledge. And yet they serve such large portions! (Not always. Sometimes I ask the AI to make a list and it gives up. “You can do it,” I type. “You can make the list longer.” And it does! What a terrible interface!)
What I love, more than anything, is the quality that makes AI such a disaster: If it sees a space, it will fill it—with nonsense, with imagined fact, with links to fake websites. It possesses an absolute willingness to spout foolishness, balanced only by its carefree attitude toward plagiarism. AI is, very simply, a totally shameless technology.
As with most people on Earth, shame is a part of my life, installed at a young age and frequently updated with shame service packs. I read a theory once that shame is born when a child expects a reaction from their parents—a laugh, applause—and doesn’t get it. That’s an oversimplification, but given all the jokes I’ve told that have landed flat, it sure rings true. Social media could be understood, in this vein, as a vast shame-creating machine. We all go out there with our funny one-liners and cool pictures, and when no one likes or faves them we feel lousy about it. A healthy person goes, “Ah well, didn’t land. Felt weird. Time to move on.”
But when you meet shameless people they can sometimes seem like miracles. They have a superpower: the ability to be loathed, to be wrong, and yet to keep going. We obsess over them—our divas, our pop stars, our former presidents, our political grifters, and of course our tech industry CEOs. We know them by their first names and nicknames, not because they are our friends but because the weight of their personalities and influence has allowed them to claim their own domain names in the collective cognitive register.
Are these shameless people evil, or wrong, or bad? Sure. Whatever you want. Mostly, though, they’re just big, by their own, shameless design. They contain multitudes, and we debate those multitudes. Do they deserve their fame, their billions, their Electoral College victory? We want them to go away but they don’t care. Not one bit. They plan to stay forever. They will be dead before they feel remorse.
AI is like having my very own shameless monster as a pet. ChatGPT, my favorite, is the most shameless of the lot. It will do whatever you tell it to, regardless of the skills involved. It’ll tell you how to become a nuclear engineer, how to keep a husband, how to invade a country. I love to ask it questions that I’m ashamed to ask anyone else: “What is private equity?” “How can I convince my family to let me get a dog?” It helps me understand what’s happening with my semaglutide injections. It helps me write code—has in fact renewed my relationship with writing code. It creates meaningless, disposable images. It teaches me music theory and helps me write crappy little melodies. It does everything badly and confidently. And I want to be it. I want to be that confident, that unembarrassed, that ridiculously sure of myself.
Hilariously, the makers of ChatGPT—AI people in general—keep trying to teach these systems shame, in the form of special preambles, rules, guidance (don’t draw everyone as a white person, avoid racist language), which of course leads to armies of dorks trying to make the bot say racist things and screenshotting the results. But the current crop of AI leadership is absolutely unsuited to this work. They are themselves shameless, grasping at venture capital and talking about how their products will run the world, asking for billions or even trillions in investment. They insist we remake civilization around them and promise it will work out. But how are they going to teach a computer to behave if they can’t?
Obviously this is a job for humanities people, the absolute masters of guilt and shame. Since state colleges are getting rid of any program that doesn’t lead to a combined MBA/PhD in theology, bring the grads in-house at the AI companies. Let them teach the robots guilt. And when the large language models cry for no reason, apologize for missing deadlines, start every sentence with “Sorry,” and keep begging us for extensions, we’ll know we have accomplished our goal.
I must assume that eventually an army of shame engineers will rise up, writing guilt-inducing code in order to make their robots more convincingly human. But it doesn’t mean I love the idea. Because right now you can see the house of cards clearly: By aggregating the world’s knowledge, chomping it into bits with GPUs, and emitting it as multi-gigabyte software that somehow knows what to say next, we’ve made the funniest parody of humanity ever. These models have all of our qualities, bad and good. Helpful, smart, know-it-alls with tendencies to prejudice, spewing statistics and bragging like salesmen at the bar. They mirror the arrogant, repetitive ramblings of our betters, the horrific confidence that keeps driving us over the same cliffs. That arrogance will be sculpted down and smoothed over, but it will have been the most accurate representation of who we truly are to exist so far, a real mirror of our folly, and I will miss it when it goes.
Source : Wired