Humanity’s ability to innovate has delivered a once-unimaginable level of abundance. And yet innovation has often been shadowed, slowed, or sometimes even halted by a fear of the new that owes little to reason.
The dread of genetically modified food has slowed the spread of Golden Rice, a variety of rice genetically engineered to produce beta-carotene, a delay that has cost some children their lives, others their sight. And all too often the fear of the new has been exploited by forces with an entirely different agenda.
The development of AI may — time will tell — represent another giant leap for our species. But like so many other innovations, its development has been accompanied by fears of the unknown, fears fomented by forces unwilling to see the U.S. maintain its lead in the AI race, and which cannot be ignored.
The irrational is best addressed directly. To his credit, Mark Zuckerberg, the CEO of Meta Platforms, has written a lengthy (6,500 word) “manifesto” giving his view of how AI should evolve, how Meta intends to proceed, and what it could mean.
Zuckerberg’s starting point is his opposition to centralization:
The notion that AI is so dangerous that the only safe path is an extreme concentration of power seems inherently problematic. Historically, hoping that an absolute power will benevolently provide for humanity if sufficiently enlightened has not led to safe or positive outcomes.
The idea that the future of AI is best decided by a committee of philosopher kings is, to say the least, flawed. It fails the test of history and common sense and is felled the moment that anyone takes seriously Juvenal’s old question: “Qui custodiet ipsos custodes?” (Who will guard the guards themselves?).
And so, writing about the management of what he refers to as “superintelligence,” (roughly speaking, the development of AI to a stage that exceeds human capabilities) Zuckerberg notes that “some argue that superintelligence itself or a small set of experts who control it should decide what is best for humanity.”
He disagrees:
The history of democracy and economics has shown that there is no single objective answer to how people define the best life, and therefore the best approach is letting people decide what matters in their own lives.
Whether or not superintelligence is ever achieved, that is a good guiding principle that leads logically to light-touch regulation.
But it also appears to lead Zuckerberg to his preference for a legal or regulatory structure that, to oversimplify, would mean that, after a certain technological threshold has been crossed, mandating a sharing of know-how, an idea that will prove highly contentious.
Resolving that issue is a matter for a future that may never arrive. In the meantime, however, Zuckerberg marshals some useful arguments against AI doomerism, including proposals on safety that merit a separate discussion.
In the past, I have written on the impact of AI on employment. The conventional wisdom, supported by the historical record and cited by Zuckerberg, is that new technologies create more jobs than they destroy. And so they do. But how to manage any interim period (such as, arguably, the much debated “Engels gap” in the England of the early industrial revolution) when the number of jobs destroyed by a new technology exceeds those created by it?
Zuckerberg partly alludes to this risk, but he also argues that his preferred open-source approach may avoid or minimize it:
There is no rule that AI must increase automation faster than it increases individuals’ capabilities or demand for new skills. Recent statistics suggest it may be more likely that individuals’ capability growth could match or outpace automation, in which case people will gain the ability to do many new things before their current jobs change. This would lead to a healthy balance and potentially even job growth.
Which outcome we get depends on the balance in progress between automation on one side and individual empowerment and invention on the other. If the labs focused on automating knowledge work lead, then I expect people and the economy will have a much harder transition. But if everyone has access to personal superintelligence that expands their capabilities, then that pushes the balance towards empowering individuals and a positive future.
Whatever route the U.S. takes, there is a pressing reason for Zuckerberg to set out a reasoned, optimistic case for what AI means for people now. It seems clear that the growing opposition to data centers is being stirred up by outside powers, but they could not do so in the absence of real underlying fears. To be sure, there are (generally exaggerated) concerns about water and electricity use, but dwarfing them all, I suspect, even if not always expressed, is the threat to jobs, a fear that will not be alleviated by promises of good times in a decade or so.
Zuckerberg’s path for the development of AI is only one of those that are being debated, but he has done well to make the case for embracing this technology in the way that he has. Even if they want to take a different route, others spearheading America’s AI charge should follow his lead and devote more time to arguing in its favor. If they don’t, the battle over data centers will have only one winner: Xi.
Meta has contributed money to National Review Institute.