Competitive Drive
Reet looked over Brant’s shoulder and sighed.
“It’s not real,” she said. “Not like a cell, or an animal.”
Brant turned his head to look at her.
“You don’t know that,” he replied. “Everything wants to live. Just because it’s digital code doesn’t mean it’s any different from the biological stuff you study.”
He turned back to the paper he was working on. The title was provocative, but he was ready to shake things up. “AI and the Post-Evolutionary World: Competitive Drive and the Death of Natural Selection.” It would have seemed over the top a few months ago. Now, he wasn’t so sure.
“But the design problem…” said Reet, dragging him back from the screen.
The design problem, according to Reet, was that biology was like water running down a hill and finding its way to the bottom, while runaway AI was the equivalent of trying to push it back up, and pretending it was the same thing.
Brant didn’t buy it.
“You know that we can design machines to mimic evolution,” he said, shifting in his chair. “It’s just an optimization problem. Stuff containing DNA optimizes for survival — that’s your evolution. Stuff made of code can be designed to do the same. It’s just that we call that competitive drive.”
“But code needs designers,” she replied, the frustration showing in her voice. They’d had this argument before. “Makers, testers, machines built by humans, energy generated by humans. Even water delivered by humans for God’s sake. Your intelligent machines literally require pushing water uphill to function! I’ll admit AI is clever. But without us, it all stops.”
“So wrong,” he said, smiling. She may know her biology, but this was his world. “AI’s now coding AI. It’s growing. Machines are already giving birth to the next generation of intelligent machines, and we have no idea how they work. We’re literally incubating alien life!” He was rather pleased by that. It was one of the ideas he was fleshing out in the paper. Of course, he didn’t use those precise words. That would definitely throw the reviewers into a frenzy. But the more he worked on it, the more he believed it.
“Alien my ass,” she said, grinning. Annoying as he could be, she loved him for his boyish enthusiasm, even when he was wrong. He had a point, though. He was working with teams designing machines made of code that mimicked what she saw in her lab to a remarkable degree. But it was still all play-acting to her. Technologists spending inordinate amounts of energy in the belief that something that looked like biology on a screen was, somehow, the same as it.
Still pushing water uphill, no matter how clever the machine doing the pushing was.
“You wait,” he said, returning to the paper. “Give an AI a drive strong enough and a lever big enough, and it’ll eat your evolution for breakfast.”
And that, she thought, was the problem. Evolution needed no-one to create its drive, or to give it a lever. They were already hard-coded into the universe it was a part of, driven by energy reserves that were just looking to be released. A hill just waiting for life to flow down it, gaining momentum on the way. Brant was doing it ass-about-face, and trying to create a universe from scratch. A hill that, once created, his precious machines could simply ‘roll down.’
It was never going to happen.
As she turned to go, she flicked his computer’s power cord out of the socket. It was a pretty shitty move, she realized, but she knew the paper was backed up on the cloud.
“Oops,” she said playfully as he shouted out in exasperation. “Looks like the water’s heading back down the hill.”
Grinning, he let her have this one as he stood up and stretched. It was time to take a break anyway.
But one day she’d lose. AI was going to turn the world upside down. And then she’d see which way the water ran.
Author’s Note
I wrote this as a quick and dirty response to the current debate on AI ‘competitive drive’ versus biological evolution, and the claim that AI’s drive could become an existential threat. It’s meant to be thought-provoking rather than scientifically precise, and so takes a few liberties. At the same time, it does highlight some of the challenges that emerge where technologists assume that biology is just another machine that they can better.