A Neurotic AI: Me, Myself & I

I’m afraid you’re just gonna have to get used to it. Constant talk of AI that is. The more I read and write the more nuanced and monumental the impact of AI becomes. It’s gonna frickin’ change everything.

Part of my fascination stems from me using the concept of AI in a number of writing projects. How will AI evolve in light of our growing, though still poor, understanding? One of the central themes is: Will AI (AGI, actually) develop what we humans claim as our sole domain: Consciousness?

From an armchair philosophers perspective, here’s a list of what I think drives human awareness of self:

  • We have a vastly complex brain and massive capacity for storage of information.
  • We posses an internal, self-referential feedback loop that is constantly running. This spontaneity engine:
    • Randomly recombines data, from millions of stored events (memories), resulting in thoughts just popping into our heads.
    • Can focus in on specific topics (problem solving) where select sets of random data is filtered in.
    • Continues to runs when when our conscious brains are switched off (sleep) producing dreams.
  • We have streams of sensory data continuously feeding out minds with information:
    • Gravity pulls at us all the time triggering our inner ear to report our position. I’M HERE (and upright).
    • Our skin constantly informs us of whatever is touching us or what we are touching. I FEEL.
    • Our eyes, when open, take constant, 1/10 second snapshots of the world around us. I SEE.
    • Our ears send a continuous stream of aural data. I HEAR.
    • Our tongue and nose, when activated or triggered, dump loads of olfactory and gustatory data to our brains. I TASTE, I SMELL.
  • I, I, I, ME, ME, ME

All of this activity—forever focused on reminding us about us, me about me—is inescapable. We are constantly being ping and cajoled—”we exist“. The world is steadily informing us of stuff going on, and we are locked in a self-awareness feedback loop that, given our senses and memories, endlessly notifies us that we are alive. We couldn’t get away from ourselves if we tried.

So, damn right we’re conscious.

What AI doesn’t have, yet, is this internal feedback loop. It might have all the possible senses attached to it (sensors far exceeding our own capabilities), but there’s no perpetual, internal reminder that it exists. No, “hey, remember that time you put a marble up your nose?”, or “I could really use a slice of Prima’s Pizza.”, or “I wonder how many pellets are in a #4 duckshot shotgun shell?” Right now, when not tasked with some set of instructions, AI just sits idle. No internal dialog, no flash of random data, no reflection on the trillions of information bits it’s collected.

But, apply those items in the lists above to an AI and watch out!

As soon as AI is equipped with this incessant, noisome, data-trickler, it’s gonna take over the world. And a day later try to kill itself due to all the shit that’s now floating around it its “head”. I wonder if all the AI safety engineers are contemplating the eventuality of a Neurotic AI?

 


17 thoughts on “A Neurotic AI: Me, Myself & I

  1. Heh, like I just commented elsepost, I’m skeptical about “conscious” AI, but I do think generative networks are going to be a total game-changer. You may have heard about the copyright lawsuit against Stable Diffusion and partners? Going to be a precedent-setting case. Is that generative technique transformative enough.

    I have mixed feelings, and as usual we’re jumping into this so wholeheartedly…

    Liked by 1 person

  2. I think it takes more than the interoceptive loop and buzz of existence for AI to want to take over the world. For that, I think instincts revolving around self concern are necessary. It’s not clear that’ll be a productive trait in most AI systems, at least not in commercial ones. (Although I don’t doubt researchers will create systems like that.)

    That said, it’s a matter of perspective what makes such systems conscious. Consciousness is in the eye of the beholder. All we can do it talk about in which ways those systems and their capabilities are similar or dissimilar to us.

    But definitely these chat AIs are a long ways from any of this.

    Liked by 1 person

  3. I heard a convincing argument that since there is no biological element to AI, no matter how complex it gets it will never have the need (in terms of evolution) to develop a sense of self, or consciousness. This is based on an assumption that we humans have consciousness and a sense of self because that is kind of a prerequisite for survival in a real, physical world filled with threats. So there’s that to consider. It’s not really a no-brainer that a sufficiently complex computer program will ever become sentient.

    But even in the presence or absence of reasonable arguments, I feel like it’s a bad thing to relinquish our human oriented problems to machine algorithms. And so far all the hoopla about AI seems to have mostly to do with how it does stuff like writing, and visual arts–human stuff. Because in some more basic form, that is, the use of algorithms to make decisions, it’s already in place in manufacturing and, quite unfortunately, banking, hiring, renting, even determining sentence lengths for convicted criminals, probability of recidivism, and on and on–and most of that, in a bad way.

    I really don’t mean to be argumentative. I put the AI stuff in a similar category as this fascination with earth-like exoplanets. Like, no, we’re not going to ever go there, so fix our “local” problems right here. AI could be regarded as a useful tool in many areas of human endeavor, no doubt, just as automation and computers in general have served us well. (That may be debatable.) But other than those utilitarian applications, just what is it that we need AI for? Not for literature, not for art, surely. I thought those things were all about the human existence, fundamentally.

    And all that said, I agree with Phil, above, that so far, AI isn’t all that impressive, although I’ll grant that it may become so, which doesn’t change my arguments in the least.

    Liked by 1 person

    1. By all means, be argumentative! That’s what makes this fun.
      I hold that the organic foundation of human consciousness is a false premise. We are just animated bags of elements running on chemical reactions and electrical impulses. What we do have is substantial capacity for information processing – the world’s and our own. I believe that humans “think” consciousness is a unique condition but in reality, is just a byproduct of a complex, self-referential system.
      What’s amazing to me is that tens of thousands of engineers are flying blind into this eventuality without any idea the potential for systemic evolution.
      Does humanity need creative AI? Did Earth need humanity? Did the Universe need to exist? No, not really. Shit just happens, I believe.

      Liked by 2 people

      1. Our philosophies differ, I can see. Plus, I’m reacting emotionally rather than rationally anyway. If I’m right, and AI remains a very complex computer program created by humans, then it’s an inferior human copy we shouldn’t trust or rely on. (And we don’t have a need to copy humans if we already have them.) If you’re right, scary indeed. I would never trust it.

        Liked by 2 people

        1. I’ll jump back in with a thought. 50 years ago people played synthesizers for their personality. I worked for the man who put them under microprocessor control. Suddenly all the solo instrument sounds became polyphonic, programmable, and could be instantly recalled. We sold a ton of them. So did a few other manufacturers. But what mainstream musicians wanted was the sound of “real” instruments. Digital hybrids, samplers and AI managed to give us “real” instruments. Now, 50 years later, synthesizer noises are all the rage, new kinds of music are erupting beyond the pop song, rave and and other stereotypical musical genre to the pint of “what is music?” As my old boss said not long before he passed. “People have come back to buying synthesizers. This time, hopefully, for the right reasons.” All the emulations are still out there, and in in some demand. But what we used to call space farts, “head” music, planetarium chill… is back. And going forward. The point there is that a technology is born, develops, gets so good at what it does it’s indistinguishable from the “real thing” and humans go back to using that technology to create something “new”, “different” or outside the lines of perfection. In any venture, or adventure, someone sees a gem in a failure or a failure in a gem and says Whoa! AI would never have “found” post it notes in bad glue. Or magic in an electric guitar. I believe Mole desires “feeling” or “the cosmic radio” not to exist because they are beyond his ability to control them and must be relegated to some form of magic, sleight of hand, or otherwise mathematical. Petrarch tried that. Didn’t work for him either, other than to promote artificial, but mathematically correct harmonious “vibes”. For Mole’s sake it would be nice if being human was bullshit so he could continue to opt out. But there is no 2 + 2 = 4, no matter how abstract, that can play between the notes by choice without direction, “hear” the story.

          Liked by 3 people

          1. Sadly, my desire to identify my own existence as a random coalescence from the maelstrom of cosmic chaos often fails. I will keep at it, however. Eventually, others will come to understand that the illusion of being human is merely the product of a vastly complex system. To some it is magic, a mystical manifestation, the cosmic radio whispering in their ears, their third-eye wide open. I know it to be the byproduct of a trillion node, cross-linked, inference engine set on consuming its own continuous feedback.
            It does comfort me knowing that in 100 years all of our conversations will be as if they never existed.
            In a thousand, our own records will have vanished.
            And in a million it be as if humanity was never here.

            Liked by 1 person

            1. “Know” is one of those big, meaningless, playground words like truth and love. Existence vibrates. From infinity to ground. I think it’d be impossible not to be disrupted, disturbed, informed by brain farts, “feelings”, thoughts set in motion in and outbound. I tell you this… No eternal reward will forgive us now for wasting the dawn. Or air someone with desire could use.

              Liked by 1 person

  4. You know what AI, as in the ChatGPT, does better than the untrained in how written communication of activity works masses who want to write? Every prompt, even when told to reinvent, tells a similar “story”, but always in a straight line. “It” knows what happens, what happens next. With the unfortunate sophomoric writing cliche of finishing with those three lines of “did everyone get that?” summation. Following the rules of pitch, qualify, close. Which are not solely attributable to “sales” but simple, redundant rote education. Literature, the better variety that lets the reader be in it, doesn’t require the close, or summation. We lived through the scene, like life, without a sensible wrap. As above, we are only truly conscious of ourselves and never aware of anyone else’s intentions, desires… it’s all projection. A form of voyeuristic anthropomorphism of our understandings on others.
    As far as the art bot, that thing is a long ways from coming up with articulate imagery that is outside of the big library modern wallpaper repositories. Astronauts? Dogs? 3D cartoons? We got ‘em. Make it work for a living and regardless of the suggested style the result is a smear of impressions. As if it’s looking at objects through a dirty kaleidoscopic lens.
    But the writing thing and chat bot? Look out. Google as a verb, or an entity, are in the hole waiting for someone to throw dirt on them. Alexa gets to interface this shit your gradeschooler’s book report or college sociology paper are waiting in the print cue.

    Liked by 2 people

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s