Bernie Sanders argues that half of the value of the AI companies should become public property and be put into a sovereign wealth fund. Unlike most of his ideas, this one has some merit and is worthy of further discussion. What are the potential pitfalls of the proposal?
I see three major counterarguments:
- THE AI DEVELOPERS WILL ARGUE THAT PUBLIC OWNERSHIP WILL CREATE A DISINCENTIVE FOR RAPID IMPROVEMENT OF THE MODELS AND WILL CAUSE US TO FALL BEHIND THE CHINESE: This position doesn’t impress me much, because half of a potentially vast pile of wealth is still a huge amount, and the CCP is going to keep tight reins on AI in order to maintain control of the country. All of the available evidence suggests that the Chinese government simply doesn’t view the AI race in the same way Trump and our capitalists do.
- GIVING TRUMP THE POWER TO CONTROL A FUND THAT COULD BE WORTH WELL OVER A TRILLION DOLLARS SOUNDS LIKE SUICIDE: The obvious way to run the fund would be to make its governing body bipartisan, but the Supreme Court has probably put an end to that kind of organization with several recent decisions. Permitting Trump, or any future president, to use the fund as a slush fund would be a disaster. The workings of the fund would have to be prescribed in great detail through legislation to avoid creating a monster.
- THIS IS A TAKING OF PROPERTY THAT PUTS ALL PRIVATE PROPERTY AT RISK OF GOVERNMENT CONFISCATION: Sanders maintains that the raw material of all AI models is the accumulated wisdom of humanity from the beginning of time, relatively little of which was added by Sam Altman and Elon Musk. That argument provides a reasonable basis for a degree of public ownership. As for the slippery slope, AI is a unique form of property that deserves a unique kind of ownership and regulation. The real question is whether John Roberts will accept that kind of logic. I have my doubts.