Luigi Zingales argues that Elon Musk is right to sue OpenAI and CEO Sam Altman, given the economic principles at stake.
Elon Musk’s outlandish behavior turns off so many people that they fail to recognize when he is right, and right he is in suing OpenAI and its CEO, Sam Altman. I am not a legal scholar, and I am not providing a legal opinion on his chances of success in his suit. I am speaking as an economist interested in ensuring that the legal system does not distort economic incentives, especially incentives to act in a pro-social way. As I will argue, the greatest benefit of Musk’s suit is to highlight major holes in our legal system regulating corporate governance and taxation that need to be fixed.
Let’s start from the beginning. In 2015, Musk co-founded and financed OpenAI, a not-for-profit corporation aimed at developing artificial intelligence in the interest of humanity. While the goal seems vague, one of the criteria stated in the article of incorporation was that OpenAI “will seek to open source technology for the public benefit when applicable.” Several years later, to fund the development of more advanced models, OpenAI created a for-profit subsidiary financed by Microsoft. Last year, OpenAI kept GPT-4’s internal design secret, violating—in Musk’s eyes—its statute. As a result, Musk sued OpenAI and Altman.
What makes the case opaque is the nature of the business and the vagueness of the terms. If, instead, Musk had donated a $50 million building to a not-for-profit to provide free education to orphan children and this not-for-profit had turned around and used the building to start a swanky (and for profit) child-care center for the wealthy, nobody would dispute Musk was in the right. After all, if we want to encourage donations to socially valuable initiatives, we need to ensure that insiders cannot divert charitable contributions to other purposes.
Unfortunately, it is much easier to write binding terms for a charitable educational initiative than for a future technology. Terms like “when applicable” make the contract challenging to enforce in a court of law. I’ll let the lawyers opine on the strengths of Musk’s case. I look at the economic case, and the economic case is clear: the public benefit intent of the initial investment has been abandoned in favor of a for-profit motive. Maybe Musk’s lawsuit is just a stunt to force the California attorney general to pursue the case. For not-for-profits, the attorney general is the only one who can bring such a case, and this case needs to be pursued for public policy reasons. First, if donors cannot trust their money will be used for the purpose for which it was raised, donations to socially valuable initiatives will dwindle. Second, suppose this case is not brought and won. In that case, it will legitimize the strategy to raise the initial funds of for-profit startups through tax-deductible donations, depleting the federal coffers and crowding out donations for socially beneficial causes, as Musk’s lawsuit makes it clear.
Yet, this case highlights an even more important weakness of our legal system: how difficult it is to create an efficient organization that does not focus on profits alone. In a for-profit company, there are many mechanisms to ensure efficiency: from stock options to proxy fights and takeovers. All these mechanisms are missing in pure not-for-profit corporations.
For this reason, a decade ago, many state legislatures (including Delaware) introduced a hybrid form: the public benefit corporation (PBC). The PBC is a for-profit corporation that aims to produce public benefits while operating sustainably. Anthropic, an OpenAI rival, chose this form. Yet, the PBC form enables but does not constrain the board to act towards the social goal stated in the article of incorporation. If the board is profit-oriented, it is all but impossible for a shareholder to sue the board and force it to pursue the social goal: managerial decisions are protected by the business judgment rule and in the absence of any conflict of interest, courts do not second-guess what the board decides. Thus, as long as the socially minded founder is still around and retains her social values, as in the Patagonia case, PBCs work well, but not when she leaves or changes her mind (as seems to be the case with Altman). As a result, a PBC statute would not have prevented what happened with OpenAI.
In many situations, a single-minded focus on profits at the expense of everything else can be dangerous, because of the externalities this pursuit can generate. To his credit, Musk had anticipated these externalities in AI and tried to set up an organization that would take them into account. Unfortunately, Altman was able to remove all these precautions. I do not know whether Musk’s suit will be able to undo Altman’s actions, but what I do know is that we need a system able to stop future Altmans from subverting donor intentions and circumventing tax law.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
