In the aftermath of Sam Altman's firing (as a Board Member and CEO) and subsequent return (at least as CEO for now), a closer examination of OpenAI's governance structure gives rise to some intriguing insights. Many commentators were swift to dismiss the not-for-profit model at the helm, deeming it inadequate for a tech powerhouse like ChatGPT-maker OpenAI, especially given its growing commercial ties with Microsoft. At Valoris Stewardship Catalysts, it is our business to examine different governance models, with a particular focus on how for-profit organizations incorporate sustainability into their decision-making. We believe it is important in the aftermath of the OpenAI kerfuffle to reaffirm the continued relevance of the not-for-profit model and point out some of the advantages of not-for-profit leadership that should not be discarded solely on basis of one high-profile, albeit flashy, example.
While acknowledging the governance challenges inherent in not-for-profits, such as the potential for cooptation of board members because of a lack of accountability to shareowners, there is an important aspect that critics these days often overlook—the underlying commitment to mission. Not-for-profit ownership of a commercial enterprise inherently embraces a mission-driven ethos, empowering companies to prioritize long-term goals and societal impact over short-term profits. For Mozilla (also a not-for-profit corporation overseeing a for-profit subsidiary), for example, this translates into ensuring an open and accessible internet to all.
In our previous roles within development finance institutions, we grappled with the governance challenges of microfinance entities, many of which started as NGOs and later encountered difficulties in maintaining their mission integrity as they subsequently transformed into fully-fledged commercial banks. A notable instance of this mission drift phenomenon occurred in the early 2000s with a microfinance institution in Lima, Peru, exemplified by its establishment of a branch in the city's most upscale shopping center.
So, you might wonder, if not-for-profits can have certain advantages in terms of long-term direction and purpose because of their governance set-up, why did OpenAI experience a seemingly disastrous outcome? While the exact reasons behind some board members' decision to terminate Sam Altman still remain unclear, it is likely they perceived a threat to the overall mission of OpenAI and the benefits of its AI technology. The lack of transparent communication may have been intentional, possibly for legal reasons or to not ruin Mr. Altman's career entirely. It would be hasty to assert that the board entirely mismanaged the process without a fuller understanding of the circumstances.
One failure in securing the mission in this case may be OpenAI’s overly broad definition of the mission itself (i.e., "save humanity from the downsides of AI"). How does a board member operationalize their duty to such a vague, if laudable, mission? What sort of accountability can there be without a more defined set of core stakeholder groups. Not-for-profit boards, like any other, are at the end of the day not immune to the risk of mission drift. As a result, we do not recommend adopting a not-for-profit governance model for for-profit entities. Instead, firms that initially began with that model may choose to sustain it, ensuring a clearly defined and manageable mission and stakeholders that keep the board accountable.
Disclaimer: The author of this blog and his firm maintain a neutral stance in the ongoing debate over the dismissal of Altman from OpenAI's Board. This piece is intended solely to express the author’s thoughts on the not-for-profit governance model implemented at OpenAI and many other organizations.