Mark Zuckerberg is entirely right to push for easing access to US open-source AI models — not least because restrictions hand even the American market to Chinese AIs.
The Meta CEO’s larger points in the manifesto he dropped on Monday also ring generally true: All humanity will be far better served with “balance of power as the foundation of safety” rather than pretending some central authority can guarantee AI will be good for us.
As he mildly notes: “Historically, hoping that an absolute power will benevolently provide for humanity if sufficiently enlightened has not led to safe or positive outcomes.”
And let’s be clear: Sen. Bernie Sanders’ demand that US firms “ AI development” is insane, and completely contrary to “the interests of humanity.”
As Sanders ought to know, China won’t stop, and its AIs are guaranteed to be designed to serve the Chinese Communist Party and its tyrannical rule — and its dominate-the-world ambitions.
This is all the more reason for the feds to heed Zuckerberg’s call to loosen the regulations that now hold US labs back: The West absolutely must “lead the open source AI ecosystem that will make up a large percent of global AI use.”
As it stands, a lot of American companies are now using Chinese open-source models, because they’re cheap and freely accessible: It’s just insane to let Beijing beat us on our own turf.
This is not to ratify every iota of Zuckerberg’s agenda: We still believe, for example, that AI developers should pay to train their models on other people’s intellectual property — though AI companies can certainly agree to allow “distilling” of each others’ models.
Above all, his central vision is compelling: “A philosophy based on individual empowerment as the source of prosperity” is the exact opposite of the CCP’s approach to everything, and the essence of the American way.