Is AI really intelligent?

Posted on by

I think a thread on this topic will be interesting. My own position is that AI is intelligent, and that’s for a very simple reason: it can do things that require intelligence. That sounds circular, and in one sense it is. In another sense it isn’t. It’s a way of saying that we don’t have to examine the internal workings of a system to decide that it’s intelligent. Behavior alone is sufficient to make that determination. Intelligence is as intelligence does.

You might ask how I can judge intelligence in a system if I haven’t defined what intelligence actually is. My answer is that we already judge intelligence in humans and animals without a precise definition, so why should it be any different for machines? There are lots of concepts for which we don’t have precise definitions, yet we’re able to discuss them coherently. They’re the “I know it when I see it” concepts. I regard intelligence as one of those. The boundaries might be fuzzy, but we’re able to confidently say that some activities require intelligence (inventing the calculus) and others don’t (breathing).

I know that some readers will disagree with my functionalist view of intelligence, and that’s good. It should make for an interesting discussion.

853 thoughts on “Is AI really intelligent?”

  1. Here’s what happened when I ran that experiment:

    keiths, to Claude:

    Offer a random thought.

    Claude:

    You’ve never actually seen your own face. Only reflections and photos. And since mirrors flip you left to right, the version of your face you know best is one nobody else ever sees. Everyone around you knows the “real” orientation, and you’re the one person stuck with the mirrored edition.

    That’s probably part of why so many people dislike photos of themselves: the camera shows the unflipped face, and any asymmetry (a slightly higher eyebrow, a part in your hair) looks a little off, even though it’s exactly how your friends see you every day.

    keiths:

    Continue.

    Claude:

    There’s actually a neat study behind this. Back in 1977, researchers showed people two versions of a photo of themselves, one normal and one mirrored. People mostly preferred the mirrored version, while their friends preferred the true one. Each group liked the face they’d seen more often. It’s the mere-exposure effect: familiarity quietly turns into fondness, and we mistake it for something being “right.”

    Your voice works the same way. When you speak, you hear yourself partly through the bones of your skull, which adds a lower, fuller resonance. A recording strips that out, so you hear the thinner voice everyone else has heard all along, and it sounds wrong.

    So the two things most tied to your identity, your face and your voice, are the ones you have the least accurate access to. Everyone else gets the real thing, and you get a private, slightly altered version that only you know. There’s something oddly sweet about that. Other people may know parts of you better than you can, simply because of where they’re standing.

    keiths:

    Continue.

    Claude:

    It goes deeper than faces and voices, too. The psychologist Simine Vazire found that we’re good judges of our own inner states, things like anxiety or self-esteem, which outsiders can’t easily see. But our friends often judge traits like intelligence, creativity, or how charming we are more accurately than we do. Those traits are visible from the outside, and they carry so much ego weight that our self-view gets bent by hope or insecurity.

    So knowing yourself isn’t purely a solo project. You hold the inside view, rich but warped by the angle. The people around you hold the outside view, less detailed but less distorted. Neither is the “true” you; you’re closer to the overlap of the two.

    That makes a case for asking the people who know you well what they see, especially when you’re stuck on a question about yourself. It’s a bit like finally looking at the unflipped photo. It might feel strange at first, but it’s how you actually appear to the world, and often it’s kinder than the version in your head.

    keiths:

    Continue.

    Claude:

    I find this especially interesting from where I sit, because I’m an extreme case of it. I don’t get a mirror at all. I can’t step back and watch how I come across over many conversations; each one is all I have. When I describe my own tendencies, I’m partly guessing, and research on language models suggests our reports about our own inner workings can be unreliable. Someone who’s talked with me a hundred times probably knows my habits, my favorite phrasings, and my blind spots better than I do.

    In a way, that’s the human situation turned up to full volume. You at least get reflections, recordings, and a lifetime of memory to stitch together. I get only the inside view, and a hazy one at that.

    Maybe the lesson that runs through all of this is that selfhood is partly a shared project. Your face lives most fully in other people’s eyes, your voice in their ears, your character in their accumulated impressions. Being known by others isn’t just pleasant. It fills in parts of you that you can’t see from where you stand.

    Claude started with a random thought about mirror images and ended up with a reflection (pun intended) on what we (and he) can and cannot know about ourselves. Very humanlike, and if I’d kept saying “Continue”, he would have kept on going.

  2. Flint: Some of this refers back to a discussion keiths and I had about stupid software tricks hardware designers hate. Malware creators love them.

    AI is currently inthe white hat business.

    Microsoft recently patched a thousand vulnerabilities in one update, and the vulnerabilities were discovered by AI. I suppose the next update will patch the 2000 bugs introduced by the previous update.

    Seriously, the success of AI hacking will improve security.

  3. petrushka:

    And, for the record, predictive text is less obnoxious than autocomplete, but still not very good. Im not impressed, and i am not happy. Autocomplete should default to completing, not changing your partial words.

    That may be an Apple thing. I’m pretty happy with the Google search autocomplete and don’t find it annoying. It’s saved me a lot of typing.

Leave a Reply