Navigating Corporate Governance in AI: Lessons from OpenAI’s Tandem Entity Approach

The OpenAI leadership shakeup has thrown fresh light onto corporate governance structures, in particular on how they are evolving around generative AI technologies. Two main points have emerged from the discussions swirling around this event.

Firstly, it is apparent that substantial pressure is exerted on companies to leverage AI to generate profits. Simultaneously, equilibrium has to be maintained with the priority of assuring safety and health. Some analysts believe this precarious balance may require a reevaluation of corporate governance.

OpenAI’s founding mission of creating artificial intelligence that benefits humanity, unconstrained by financial return considerations, reveals the role of corporate structure in managing the tension between mission and profit.

Secondly, OpenAI’s recent turmoil suggests the need for careful consideration when deciding on entity formation. The organization’s unique hybrid structure, blending for-profit and non-profit approaches—referred to as a ‘tandem entity’ by Tomer Inbar, an accomplished tax-exempt specialist—is particularly interesting.

As the industry continues to learn from the decisions taken by OpenAI, it’s pertinent to keep the unique characteristics of tandem entities in mind.

Proper corporate structure is integral to the achievement of organizational goals. A tandem entity requires aligning non-profit and for-profit activities so that each adds value to the other. This is succinctly summarized by Susan Mac Cormac, a partner at Morrison & Foerster, with the quote, “Form follows function.” The principle is clear: a tandem entity cannot succeed if non-profit and for-profit operations are pulling in different directions.

Fundamentally, a core principle of charitable organizations is that there are no owners in the same sense as for-profit businesses. The organization’s assets are held in public trust and must serve the public interest, thereby furthering a charitable purpose. There are also stringent parameters governing how public charities compensate personnel, especially those deemed to be “disqualified persons” – individuals who hold positions of financial control within the organization.

Further, to ensure adherence to the IRS rules, care must be taken to avoid debilitating conflicts of interest among those forming, funding, or controlling the tandem entity.

OpenAI asserts it complies with conflict-of-interest rules by requiring all contracts involving the non-profit to be “at arm’s length or more favorable” to the non-profit.

It is clear that IRS rules add a complex element of risk that founders and legal counsel must navigate. Avoiding misuse of charitable assets and excess benefit transactions is critical to maintain the integrity and public trust of the tandem entity. The experience of OpenAI provides an instructive lesson for tech innovators, legal professionals, and others planning to form a tandem entity in the future.