Schneier Calls For US Nationalization Of AI Giants

Someone is being paid for the mere fact of ownership - for holding equity in a company whose product, by the admission of its own architects, may pose risks to the public that no market signal is competent to price. What service, precisely, does this payment purchase for the society that provides it? Bruce Schneier and Nathan E Sanders, in proposing that the United States government stand ready to nationalize firms such as OpenAI and Anthropic should the markets reject them, have put their finger, whether they know it or not, on the oldest question in the moral economy of property: does this claim of ownership correspond to a function performed, or does it survive only because no one has yet asked what it is for?

The nineteenth-century mill owner justified his profit, however brutally, by pointing to the mill: the coal moved, the cloth was woven, something was made that the world could use and see. The shareholder in an artificial intelligence company justifies his claim by pointing to something else - a model, a valuation, a projected capability - the value of which rests substantially on the expectation that others will believe in it. This is not a new arrangement in itself; finance has always dealt in expectation. What is new, and what Schneier and Sanders have grasped, is that the thing being speculated upon is not a bridge or a railway but the cognitive infrastructure on which a society may come to depend for medicine, education, warfare, and governance, and that its builders have said openly that it might fail catastrophically and that markets are not equipped to prevent this.

Here the functionless wealth test cuts cleanly. If the profits of OpenAI and Anthropic represent a fair return on a service reliably rendered - models that inform, that assist, that extend human capacity without corrupting the civic fabric that surrounds them - then their shareholders have earned their claim as the ironmonger earned his. But if those profits instead represent a wager on capability racing ahead of safety, sustained by capital markets whose only discipline is the fear of being left behind by Google and Meta, then what is being rewarded is not function but velocity. The corporation is no longer an economic institution answerable for what it does; it has become, in the old and precise sense, an engine of accumulation answerable to nothing but its own continuation.

Nationalization, on this reckoning, is not a punishment but a restoration of the question that acquisitive enterprise prefers to leave unasked. To say that the state should assume ownership if the markets reject these firms is to say that the function - safe, broadly beneficial artificial intelligence - must not be permitted to die merely because the private claim upon it has failed to pay. This is the argument, at bottom, of every serious socialist critique of the joint-stock company: that ownership was always meant to be a trustee’s office and has decayed, gradually and almost invisibly, into a licence to collect. The mistake of the libertarian objection - that such a proposal is merely a seizure dressed as prudence - is to suppose that private ownership under current conditions represents any function at all beyond risk-bearing for its own sake, when the risk in question is one the shareholder can walk away from and the society cannot.

I do not think Schneier and Sanders have solved the problem so much as located it correctly. A government competent to run a model safely is not guaranteed by the act of ownership any more than a private board is guaranteed competence by the act of raising capital; the state, too, must be asked what function it performs, and by whom it is watched. But the proposal has the merit of refusing to treat the market’s verdict as morally final. A man who owns a share in a great AI company and does nothing but wait for its price to rise is not a builder. He is a claimant. And a civilisation that allows the architecture of its own thought to be owned by claimants, rather than governed by those who can be made to answer for its function, has arranged its affairs very foolishly indeed.