The Difference Between Optimizing and Evolving
Artificial intelligences are not alive.
They do not evolve. They may iterate and optimize, but that is not evolution. Evolution is random mutation in a particular environment. Machine learning, by contrast, is directed toward particular, preprogrammed ends. It may be complicated, but — unlike evolution, the weather, the oceans, or nature — it’s not complex. Complicated systems, such as a city’s many traffic lights, direct activity from the top down. By contrast, complex systems such as a traffic circle establish flows spontaneously through the interaction of their many participants. Machines have lots of complicated parts and processes, but no higher order; lifelike complexity is emerging from them.
No matter how many neurological-sounding terms are coined for what computers do, they are not on the path to consciousness. Those fantasizing about computer life like to throw around terms like “fuzzy logic,” as if these programming techniques are bringing machines closer to exhibiting human intuition. Fuzzy logic is the ability of a computer program to consider values other than one or zero, and then express them as a one or a zero. That’s all. Fuzzy logic isn’t fuzzy in the manner of true uncertainty. It merely reduces the roughness and complexity of reality into a simple binary that a computer can work with.
Similarly, neural nets are not like human brains. They’re simply layers of nodes that learn to do things by being fed hundreds or thousands of examples. Instead of telling the computer what a cat looks like, we just feed it hundreds of pictures until it can determine the common, distinguishing features. Human brains are capable of generalizing a category like “cat” after see-ing just one example. How? We’re not exactly sure.
Our inability to say exactly what it means to be a thinking, autonomous human being should not be considered a liability. The human mind is not computational, any more than reality itself is just information. Intelligence is a fabulous capability of the brain, and reality stores immense quantities of data — but neither of these exists at all without human consciousness to render them. We must not reduce human awareness to raw processing power. That’s akin to reducing the human body to weight lifting. Our calculation speeds can’t compete with those of a supercomputer, and we will never lift as much as a crane. But humans are worth more than our utility. Enhancing one job-friendly metric with technological intervention or replacement just leaves other, likely more important values behind. The most important of those is consciousness itself.
As best we know, consciousness is based on totally non-computable quantum states in the tiniest structures of the brain, called microtubules. There are so many billions of these microtubules, and then so many active, vibrating sites on each one, that a machine harnessing every computer chip ever made would wither under the complexity of one human brain.
The only people who behave as though consciousness were simple enough to be replicated by machine are computer developers. Real neuroscientists remain delightfully flummoxed by the improbability of self-awareness emanating from a lump of neurons. It’s confounding and paradoxical.
That doesn’t mean we will soon be able to dismiss consciousness. It’s not some illusion perpetrated by DNA on the human brain in order to compel a survival instinct in its hosts. We do not live in a simulation; our awareness is real. When pressed, even physicists accept that consciousness has a better claim to existence than objective reality. Quantum theory holds that objective reality may not exist until we observe it. In other words, the universe is a bunch of possibilities until someone’s consciousness arrives on the scene and sees it a certain way. Then it condenses into what we call reality.
The search for the seed of consciousness is a bit like the quest for the smallest cosmological particle. It’s more an artifact of our mechanistic science than a reflection of how the mind works. Whenever we discover one ultimate determinant—such as the gene—we also discover that its expression is determined by something else. No sooner do we find the germ we deem responsible for a disease than we also find the environmental factors allowing it to flourish, or the immune deficiency responsible for its transformation from helpful bacteria to invading pathogen. The only way to solve consciousness is through firsthand experience and reverence for the world in which we live, and the other people with whom we share it.
In this sense, we know consciousness exists because we know what it feels like. Just like an animal or a computer, we can see a cup of coffee sitting on the kitchen table. But we humans also know what it is like to see a cup of coffee on the table. The choice to look at that cup, to pay attention, is unique to consciousness. Computers can’t do that. They have to see everything in range. They have no attention. No focus. No real direction.
And to know what it is like to look at a cup of coffee, to be able to consider that construction of mind and self: Only humans can do that. That’s because we’re alive, and computers are not.