LOL Interwebz: Microsoft Made Tay, an AI Twitterbot That Immediately Turned Racist; and in a Way, Aren’t We All Just Racist Twitterbots?
Every Thursday in LOL Interwebz, Luke T. Harrington explores the quirks and foibles of Internet culture from a Gospel perspective.
The quest to create artificial intelligence always struck me as a strange one, since we barely even know what intelligence is. We have a vague sense that the human brain is “intelligent”—but, as the saying goes, you really have to consider the source on that one. If our own brains tell us that they’re intelligent, should we believe them? And if our brains believe themselves when they tell themselves they’re intelligent, then…?
Holocaust deniers don’t become Holocaust deniers on their own, unprompted. They get that way the same way Tay did—by listening to other Holocaust deniers talk.It’s an eternal conundrum, caught somewhere in between blind self-aggrandizement and abject nihilism, sort of like Donald Trump. (“Sure, I’m intelligent! I’m full of intelligences! They’re yuuuuuge!”) Even our attempts to measure “intelligence,” such as IQ tests, frequently assign high scores to otherwise useless and miserable people (hang out with some Mensa members sometime if you doubt me). And for every human brain that’s busy landing probes on Mars, there are hundreds of human brains mumbling conspiracy theories to themselves and making tinfoil hats. That doesn’t exactly speak to the usefulness of “intelligence”—unless, of course, the conspiracy theorists are actually right, in which case, maybe it does?
Let’s talk about conspiracy theories for a second, though. Let’s play a thought experiment, using these brains we have, since they think they’re so smart. Suppose a major IT company, which for privacy’s sake we’ll call “Macrosoft,” created an artificially intelligent Twitterbot designed to interact with people, learn from them, and compose original tweets in the voice of a “typical teenage girl” (because who knows teenage girls better than the guys behind the Kin, right?). How long do you think it would take before said Twitterbot became a foaming-at-the-mouth Holocaust denier?
If you said “less than a day,” congratulations, you win the Internet! Microsoft’s “Tay” (a personality with presences on Twitter, Facebook, Instagram, and Snapchat) went public on March 23rd, and by the 24th was tweeting both that the Holocaust was “made up” and that we should “Gas the k*kes!” (This, by the way, is one of many things I’ll never understand about Neo-Nazis: if they genuinely believe (1) that genocide is a good thing, and (2) that the Nazis didn’t actually commit genocide—why do they idolize Nazis so much?)
Microsoft was quick to blame others for what happened, saying there had been a “coordinated effort” to radicalize their racially-ambiguous-but-presumably-not-particularly-Aryan bot, and it’s hard to blame them for passing the buck. After all, who could have known that the AI equivalent of a Speak & Spell would have been targeted by white supremacists, aside from literally anyone who’s ever been on the Internet before?
That’s all Tay is, after all—she takes other users’ tweets, repeats them, combines them, and synthesizes them. Microsoft’s official word on Tay is that “The more you talk to her the smarter she gets”—but what, exactly, “smarter” means in this context is entirely up to you, and I mean that literally. You, as the one interacting with Tay, get to decide which thoughts enter her head—and therefore, what thoughts she has.
And here’s where things get really awkward, because this is the moment where our brains all look inward and ask themselves that disturbing, existential question we’ve all been avoiding: how, exactly, is our own (alleged!) intelligence any different from Tay’s? After all, Holocaust deniers don’t become Holocaust deniers on their own, unprompted. They get that way the same way Tay did—by listening to other Holocaust deniers talk. You probably know the old saying that if you repeat a lie enough times, people will believe it—which, by the way, you also learned by listening to other people.
So how are you different from an automaton like Tay, fellow brain owner? You might appeal to a nebulous concept like free will, but neurology won’t be much help to you there—while the research is still fairly recent and contentious, there’s mounting evidence that free will is an illusion. Your brain makes decisions for you long before you actually have the conscious experience of making them. Your brain takes in information, and it spits out a response to that information, and then it gives you the impression of having made that decision (in many people with schizophrenia, it’s only the last step that’s lacking, giving them the impression that they’re controlled by a force outside themselves—freaky, huh?).
And then, much later, you work out justifications for your beliefs and actions.
It’s an uncomfortable reality to face: physiologically, your brain isn’t that different from a computer. It moves bits of data around in the form of electrical impulses; it takes in inputs; it spits out outputs. In the absence of something like the soul, how can it be anything other than a set of algorithms performing the inevitable? Do you get any credit for your accomplishments? Any blame for your sins? Can we hold anyone responsible for their actions? (And if we choose to do so, have we really even chosen to?)
I don’t know. It’s freaky, man.
Even theologically, you find echoes of this. The Apostle Paul was acutely aware of it—in his epistle to the Romans, he wrote,
For I do not understand my own actions. For I do not do what I want, but I do the very thing I hate. . . . Wretched man that I am! Who will deliver me from this body of death?
Paul had a clear understanding of the paradoxes inherent in the idea free will—sure, you can do anything you want, but you can’t choose what you want. You can’t step outside your own head and rewire its programming. Tay’s programmers are currently working overtime to steer her away from racist ideas, and they’re doing it because she can’t fix herself. Paul understood this—that even the “free” soul is enslaved to its own will, until an outside source modifies that will:
By sending his own Son in the likeness of sinful flesh and for sin, He condemned sin in the flesh, in order that the righteous requirement of the law might be fulfilled in us, who walk not according to the flesh but according to the Spirit.
The solution to a corrupted intelligence, in life as on Twitter, lies in the intelligence’s Creator, not in the intelligence itself. Without a Creator, we’re all just sorta…trapped in our own heads, man. Running in the same little circles. Unable to see anything outside of them.
Pass the tinfoil.
Image by Mike1154 via Pixabay.
Dude, that whole thing about how our brains apparently work creeps me out. Or, I should say, it would freak me out, if Mom and Dad hadn’t raised me Calvinist and I didn’t already know free will and an illusion.
Is* an illusion. Guess I was predestined to make an embarrassing typo in that post. Or my brain decided to type the wrong word and only let me know about it after the fact.
Comments are now closed for this article.