You can’t go on a single web page without AI getting in your face, either it or something about it and how it’s going to doom us all, stocks are tanking, jobs are disappearing. But the thing is, the only people who care about AI are the ones shoveling it into everything. The CEOs and corporate executives who are investing so much into it because they took a pass on crypto and don’t want to miss the boat a second time.
Who’s going to be the 800-lb. gorilla first? Who’s going to go under because of it? How will you be able to tell what’s AI-generated and what’s not? When’s it going to replace homework? Will students now be trained on how to prompt AI instead of writing essays (the same way they told us we wouldn’t have calculators in our pocket as adults)? How are teachers and artists going to detect plagiarism? What happens when the computer makes a deep fake picture of you sucking someone’s dick? Will humanity be rendered useless? Is it going to Skynet us?
The experts in the field seem to be telling us that it’s not a matter of “if” but “when” it becomes intelligent. And we’re talking this year. They’re saying there’s going to be an entity out there that’s indistinguishable from a human, like Mega Man or Johnny 5 or V.I.N.CENT or Data. And how are we going to tell when it does? How do we define intelligence? Is it going to have rights? Does it have to have emotions? What’s the difference between having emotions and pretending to have emotions? Does it have a soul? How would we know?
Personally, I’m not worried about AI. And that’s for two reasons.
Reason the First
I think, for something to be alive, it has to have desire. It has to want something. I don’t know how we’d be able to detect that, but I know desire is the root of all man’s suffering (as Confucius says). Desire is the root of all emotions. I want this. I don’t want this. I want this to stay like this. I want this to go away from me. I liked this so I’ll get more. I want a lot of this at once, then not so much. I want some of that, but none of that. Disgust, Joy, Sadness, Fear, Disgust, Anger, Anxiety, Envy, Embarrassment, Ennui, Nostalgia–all the Inside Out emotions are about wanting something or anti-wanting something.
But how will you know that it actually wants something and isn’t just faking because the code says so. Not sure. To my knowing, AI only works by flipping a bunch of switches until it gets something right. I imagine we’d know because it will ask for something which was not one of its switches to begin with.
That’s why I’m surprised no one thought Data was a sentient being in ST:TNG. Maybe values change in 2364. He didn’t have emotions, but he had desires. He wanted to be human. Even before Star Trek: Generations, he clearly demonstrated aspirations and preferences. He pursued a career in Starfleet without prompting. He styles his hair a certain way. He likes Sherlock Holmes, painting, poetry. He’s curious–he wants knowledge. He chose a cat instead of a dog, goldfish, hamster, or any other alien pet. He had friendships–people he favored over others. And all this without anyone commanding him to do so. The guy’s sapient as hell.
But today’s AI will just sit there until you prompt it with something. It will be pleased as punch* to exist for eternity until it gets asked another question and I don’t think it or us would notice a difference.
*(not really pleased–that’s just a simile)
There’s an argument to be made that you need emotions to express desire and/or you need desire to have emotions. So it’s a chicken-and-egg thing that has no real point of creation or termination, like the “no fate” speech in The Terminator. Nobody created it. Everyone just passed it on (which boggles my mind). I don’t think you need emotions to experience desire, necessarily.
Or maybe you go back further and find out what caused human beings to experience desire. Why do bad feelings exist when you do certain things and good feelings for others? And sometimes both. Plus it seems to be a little bit different for every individual. What makes one person want this and not that?
And maybe you need to go back all the way to the first unicellular animal, maybe when they first started to develop light-sensing organelles so they could tell where there was more light and where there wasn’t. It’s got to be more than just evolutionary tactics. But then that begs the question–who engineered those?
Reason the Second
No entity will be convincing as an independent intelligent being or something indistinguishable from a human or on our level until it can do one thing. No matter how far across the uncanny valley you reach, it won’t ever make it all the way past that dip unless it’s allowed to or naturally develops a certain ability.
It needs to make mistakes.
Now, what’s delightful about this is that it’s DIRECTLY counterproductive to what AI is for. Humans constantly make mistakes–you could say it’s one of our defining features. They succumb to laziness, lust, anger, misinterpretation, etc. Computers do not. Brains forget things. Computers do not. Brains get distracted. Computers do not. Humans aren’t in control of their bodies at all. Skin does random things. Fingers tick. Eyes move around. Eyelashes flicker. You can’t even pronounce one word the same way each time, not perfectly. A computer can. The only time they fail is because humans didn’t give them the right instructions or a piece of equipment failed due to physical degradation or entropy, as all matter does.
This is why I can always tell what’s CG in a movie vs. what’s actually there. At least when it comes to organic stuff. Mountains and backgrounds and objects, not so much. But Gollum, Rocket Raccoon, Paddington, Avatar, any de-aging technologies. They just move too smoothly, too flowing. There aren’t enough jerks. An arm doesn’t swing across the body in one fluid motion, because bones aren’t fluid. They have imperfections, they aren’t perfectly smooth stainless steel in the ball and socket joint. Muscles can be sore and uneven. Eyes don’t move like globes in water, they twitch erratically, they get junk in them. Some toes are shorter than others, making the gait uneven. A CG character can remain 100% motionless like it’s on pause. But a living body is never motionless, it’s always breathing or excreting or jiggling or vibrating, as if worms are crawling under the surface.
Jurassic Park is still the most realistic CG I’ve ever seen. Not the best, but the most realistic. That’s because A) they used a lot of animatronics in conjunction so that you can’t tell when it’s real and when it’s not B) they spent a lot of time perfecting the CG. (T2 comes a close second for the same reasons). Today’s movies over-reach. They create super complex characters, but there’s always a deadline and they have to cut off production at some point.
Computers are very good at doing things perfectly. But they’re terrible at doing things imperfectly. Someone’s got to program the thing to occasionally think 2+2=5, and no one will ever do that, because that runs contra to AI’s purpose.
Super Smash Brothers Brawl (Wii) had a mechanic called “tripping”, where a character would just fall 1% of the time it tried to dash or turnaround. Players hated it–why would you want the computer to randomly halt your progress. What if racecars randomly stopped 1% of the time you hit the brake? What if your computer mouse randomly ignored a click 1/100 times? Not acceptable.
And it’s not acceptable in an AI. And that’s our advantage.
