Why Are People?
Creativity cannot be mimicked.
It sounds like a sensible question: what are people? But what if one instead poses another question: why are people? It sounds strange to the ear. But I think it is the better question. It is also the opening chapter title of Richard Dawkins’ celebrated book The Selfish Gene. Because the gap between the two questions is a gulf.
When people ask, what are people?, they look to a definition of what a person is. But when one instead asks, why are people?, one looks to an explanation of what a person is. Definitions and explanations are not the same. The first is a question about what words to use, the second is about what is in fact real.
Let’s say we try to go about trying to define what a person is. One problem we cannot escape is that others will have very different definitions for what is the same thing: a person. The discussion turns into a debate about the use and misuse of language. But we all know what we really mean. After all, I am not looking to update the Oxford English Dictionary. I’m looking to explain what’s really real.
It’s also tempting to turn definitions into tests. Does this “person” pass or fail the test of this definition of a person? What if they do, but then fail a different definition of the word person? Even Alan Turing himself thought of the “Turing Test” as a mere thought experiment, and not at all a definition of what a person is. He knew mimicking conversations said nothing about inner creativity. So definitions aren’t as helpful here. They are simply convenient shorthand for a longer description of something. The word water is a nice shorthand for molecules that bond two hydrogen atoms with one oxygen atom in a very marked way.
The only way to explain why people are people is because they solve problems in a creative, unbounded way, and this cannot be explained in any other way. The progress that a person can make is never capped from within, and they will always be at the beginning of infinity. Since creative problem-solving is unbounded, it cannot be automated. If it cannot be automated, it cannot be tested. In contrast, an AI model is not creative because it is a fixed, black-box program, which can be automated and replicated. We understand the internal workings of AI well. These things are all therefore testable.
One may ask that if one cannot test for personhood, then how can we protect artificial people (AGI), once they exist?
We can protect them by using the best explanation of why people are people. Unbounded curiosity and creativity are impossible to mimic, and there is no other explanation for such impossible behaviour other than internal creativity. Artificial people will be able to do things that cannot be automated or replicated at all. Thus they are people, just like us.