CHaTGPT and Whole Hospital Joint Ventures, Part II

We all interact with AI everyday.  Talking to the bank or credit company’s customer service computer is just one example.  Of course, those are extremely primitive versions of AI.  Calling an automated teller “AI” is silly, I know.  Comparing ChatGPT (Chat means communicate and GPT stands for “Generative Pre-Trained Transformer”nonprofit OpenAI, with its $1 billion tax free endowment, is amongst the leaders of the coming robot insurgency.  But for all intents and purposes, OpenAI is in literal, if not legal, violation of IRC 501(c)(3)’s prohibition against private inurement.  The entirety of its operations — the “whole hospital” — is devoted to a limited partnership called OpenAI, LP and “part of the net earnings of which” will necessarily “inure to private shareholders.”  Here is how Open AI (the exempt nonprofit) explains it:

We want to increase our ability to raise capital while still serving our mission, and no pre-existing legal structure we know of strikes the right balance. Our solution is to create OpenAI LP as a hybrid of a for-profit and nonprofit—which we are calling a “capped-profit” company.

The fundamental idea of OpenAI LP is that investors and employees can get a capped return if we succeed at our mission, which allows us to raise investment capital and attract employees with startup-like equity. But any returns beyond that amount—and if we are successful, we expect to generate orders of magnitude more value than we’d owe to people who invest in or work at OpenAI LP—are owned by the original OpenAI Nonprofit entity.

It’s hard to really know, but on the one hand it sounds like Open AI is getting pretty good legal advice.  But even if that is true, those advisors have made what looks like me to be an unforced error.  One that I think might, as a normative matter if nothing else, disqualify its 501(c)(3) status.  The LP agreement meets all of the organizational requirements for whole hospital joint ventures stated in Revenue Ruling 98-15.  Revenue Ruling 98-15’s first concern was whether the charitable partner’s public benefit mission would take clear precedence over the noncharitable partner’s profit motive, particularly in light a of a general partner’s fiduciary duties to limited partners.  The ruling imposed all sorts of organizational requirements to ensure the nonprofit managers dominate governance and then exclusively in accordance with the charitable mission.  But the ruling is far less explicit about when too much profit sharing is too much and I think therein lies the problem.  Open AI appears to have met the organizational requirements, in form at least:

The Mission Comes First

We’ve designed OpenAI LP to put our overall mission—ensuring the creation and adoption of safe and beneficial AGI—ahead of generating returns for investors.  The mission comes first even with respect to OpenAI LP’s structure. While we are hopeful that what we describe below will work until our mission is complete, we may update our implementation as the world changes. Regardless of how the world evolves, we are committed—legally and personally—to our mission.  OpenAI LP’s primary fiduciary obligation is to advance the aims of the OpenAI Charter, and the company is controlled by OpenAI Nonprofit’s board. All investors and employees sign agreements that OpenAI LP’s obligation to the Charter always comes first, even at the expense of some or all of their financial stake.

And . . . 

Only a minority of board members are allowed to hold financial stakes in the partnership at one time. Furthermore, only board members without such stakes can vote on decisions where the interests of limited partners and OpenAI Nonprofit’s mission may conflict—including any decisions about making payouts to investors and employees.  Only a minority of board members are allowed to hold financial stakes in the partnership at one time. Furthermore, only board members without such stakes can vote on decisions where the interests of limited partners and OpenAI Nonprofit’s mission may conflict—including any decisions about making payouts to investors and employees.

I swear I can smell a lawyer a mile away.  The nerds are too busy planning the robot takeover to have read 98-15.  Somebody did, and the governance structure is straight from the dusty, dark smoke filled basement offices where Revenue Rulings are secretly hatched (I made that part up, of course, but Grisham won’t take the screenplay bait if I don’t).    

The problem, it seems to me, is that the limited partners are essentially preferred shareholders.  Like typical preferred shareholders, they are silent with regard to governance and operations, but they get the first fruits!  I’m not sure whole hospital or private inurement jurisprudence has ever addressed the preferential allocation of net revenue to private individuals.  If the partnership generates 150x net revenue, presumably the first 100x must be paid to a “private shareholder or individual.”  If the partnership generates 100x or less, it all goes to the private investors.  Is that just a bridge too far?  The existence of a cap on total possible compensation, by the way, has been considered relevant to whether private inurement results from a revenue sharing arrangement.  What is bothering me in this instance is that private shareholders or individuals eat first (up to a maximum stated amount), and only the amount left after profit takers eat their fill is devoted to scientific research.

Maybe I am overthinking the whole thing.  Maybe without much representation on the governing board, there will be little danger that public mission will be subordinated to private profit, not even as a mathematical bottom line financial matter.  If the preference does not dictate how “net profit” is defined, maybe there is no problem ultimately.  Another thing that bothers me is the agreement allowing Google an exclusive license in the IP.  And then from an efficiency standpoint, I gotta wonder whether tax exemption is even warranted when there are private investors willing to invest billions.  Won’t the market provide AGI eventually even without public investment?  If the fear is monopolization of AGI, there are better ways to preclude any one nerd from controlling us all than by public subsidy.  Finally, from an equity standpoint, is it fair to subsidize the greater income of billionaires by public subsidy?  Who wants to co-author a law review or Tax Note?

darryll jones