If we sort people only by superior intelligence, we’re sorting people by a quality few possess; we’re inevitably creating a stratified, elitist society.
We want a society run by people who are smart, yes, but who are also wise, perceptive, curious, caring, resilient, and committed to the common good.
If we can figure out how to select for people’s motivation to grow and learn across their whole lifespan, then we are sorting people by a quality that is more democratically distributed, a quality that people can control and develop, and we will end up with a fairer and more mobile society.
~David Brooks in Atlantic Monthly Dec. 2024
The above is an example of editorial page “If we can” and “we really should” nostrums/ truisms that sound great, but have yet to create utopian civilizations. Meritocracy is good, except when it isn’t. Elites are a product of Meritocracy, but maybe we can get around that. Except when we need elites.
In some discourse post 2024 US election, Elites are anyone who didn’t vote for the winning candidate, because those voters were overwhelmingly college graduates. Except of course does a college degree actually put one into the “elite”? Who is the true elite? Is there an income amount that determines that? Or is it a certain level on organizational hierarchies?
“Wise, perceptive, curious, caring, resilient, and committed to the common good.”
That sounds kind of like how we might describe alignment of AI with human values. (Although the common good is uncommonly difficult to discern as interests splinter among groups for many many different reasons.)
But say we are able to create AI that matches what Brooks says we need in our meritocracy. Would it potentially be easier to create that “AI cadre” of elite decision makers, than it would to grow them in the numbers and quality needed in our 330 million population? What if the best way to overcome “human nature” limitations is to create AI elites that don’t have those limitations?
If elites are those who have good paying stable jobs, the common take is intellectual workers are going to be the first to lose their employment as AI is deployed to do that kind of work. If so, the meritocracy would be stood on it’s head, as those doing non-intellectual jobs would seem to still have their jobs.
But what if Robots can do plumbing, carpentry, car repair, and many many services? What jobs… what employment …would be left then?
That sounds like fantasy, but according to leading lights like Sam Altman at Open AI, and Dario Amodei at Anthropic, AGI is coming somewhere between 2025 and 2027.
Not to say the actual deployment of AGI will occur seamlessly once the capability is achieved.
There will be matters of cost efficiency, and other hurdles, but it doesn’t seem too early to start wondering who the new elites will be, and who the meritocracy will anoint next. And how we can adjust to such a situation, including figuring out what education should be about.