So, What Exactly Is AI? (And Why Nobody Can Agree)

Ever notice how everyone seems to have a completely different idea of what AI actually is? You’re not imagining it.

Here’s something that might surprise you: “The people making it don’t know what AI is either. Not really,” says Chris Olah, chief scientist at Anthropic. “These are the kinds of questions that are important enough that everyone feels like they can have an opinion. I also think you can argue about this as much as you want and there’s no evidence that’s going to contradict you right now.”

That’s from a fascinating MIT Technology Review paper by William Douglas Heaven that peels back the curtain on AI’s biggest secret—nobody can agree on what it is.

The Math vs. Magic Divide

The paper opens with a stunning observation: “AI is sexy, AI is cool. AI is entrenching inequality, upending the job market, and wrecking education. AI is a theme park ride, AI is a magic trick. AI is our final invention, AI is a moral obligation. AI is the buzzword of the decade, AI is marketing jargon from 1955. AI is humanlike, AI is alien. AI is super-smart and as dumb as dirt.”

Wait, what? Yes—all of those statements are things people actually believe about AI. Simultaneously.

Here’s the technical definition: “AI is a catchall term for a set of technologies that make computers do things that are thought to require intelligence when done by people.” Simple, right?

Except it’s not. Because “what does it mean for machines to understand speech or write a sentence?” That question alone has sparked what Heaven calls “Internet nastiness, name-calling, and other not-so-petty, world-altering disagreements.”

The paper highlights a perfect example: Critics Alex Hanna and Emily Bender mockingly replace the term “AI” with “mathy math”—just lots and lots of math. Meanwhile, OpenAI CEO Sam Altman tweets about the latest GPT-4 update: “Feels like magic to me.”

As Heaven notes: “There’s a lot of road between math and magic.”

The Greatest Trick AI Ever Pulled

Here’s where it gets really interesting. Heaven reveals his own take: “Writing about GPT-3 back in 2020, I said that the greatest trick AI ever pulled was convincing the world it exists. I still think that: We are hardwired to see intelligence in things that behave in certain ways, whether it’s there or not.”

Think about that for a moment. We might be projecting human-like intelligence onto something that’s fundamentally different.

Researcher Pavlick adds: “I’ve never had to think about whether a piece of language required intelligence to generate because I’ve just never dealt with language that didn’t.” We’re in uncharted territory.

Why This Matters to You

Heaven’s conclusion is refreshingly honest: “AI is not one thing; it never has been, no matter how often the branding gets seared into the outside of the box.”

He continues: “The truth is these words—intelligence, reasoning, understanding, and more—were defined before there was a need to be really precise about it. Words get redefined and concepts evolve all the time.”

So what should we do with this? Heaven suggests: “The sooner we can all take a step back, agree on what we don’t know, and accept that none of this is yet a done deal, the sooner we can… stop calling each other names.”

The Bottom Line

AI isn’t magic. It isn’t just math. It isn’t humanlike, and it’s not the solution to all our problems. It’s “an idea, a vision, a kind of wish fulfillment” that means different things to different people—shaped by morals, worldviews, politics, and gut instinct.

And maybe that’s okay. Maybe the first step toward using AI responsibly is admitting we’re all still figuring out what it actually is.

Download the source paper: What is AI? – MIT Technology Review (PDF) https://drive.google.com/file/d/1ZY1MWD_whF4bV1EKr2cWXoREggUU3YaX/view?usp=sharing