Background
When OpenAI was created in 2015 as a non-profit research lab, its stated goal was “advancing digital intelligence in the way most likely to benefit humanity as a whole, unconstrained by a need to generate financial return.” It had no products and was funded by donations of dollars and computing infrastructure.
In 2019, when it became clear how much capital would be needed to create AI models, OpenAI switched to a novel structure where a new for-profit arm would be under the control of the non-profit board. As a symbol of its continued commitment to the public good, the for-profit arm capped the share of profit employees and investors could receive.
As the AI market exploded and investors poured hundreds of billions of dollars into it, there has been increasing tension between OpenAI’s non-profit mission and the for-profit arm’s desire to capture a share of the wealth it helped create. Concerns arose over whether OpenAI’s CEO, Sam Altman, had abandoned the goal of AI safety in favor of making money. Some non-profit board members accused him of lying to the board. He was fired but reinstated shortly thereafter.
OpenAI is now planning to transfer control of the organization from the non-profit board to a new for-profit, publicly-traded Delaware Public Benefit Corporation. The non-profit arm will be independent, but funded by shares of the for-profit company. The non-profit arm will use its resources to “pursue charitable initiatives in sectors such as health care, education, and science.”
Ostensibly, a public benefit corporation allows corporate leaders to balance employee and customer welfare with their fiduciary duty to seek profits. OpenAI competitors Anthropic and Elon Musk’s xAI also use this structure. However, the public benefit designation is mostly symbolic, as there are no rules or enforcement mechanisms to assure that the public is actually benefiting from the organization’s work. Several public benefit companies, including a well-known for-profit university, have been caught up in fraudulent or sketchy activities.
The planned transition is not a done deal. Elon Musk, who has a lawsuit pending against OpenAI for anti-competitive practices, has filed a motion to stop OpenAI from becoming a for-profit company because he claims it will prevent OpenAI from having the resources to pay should he win. Meta CEO Mark Zuckerberg has asked the California Attorney General to block the transition, arguing that the organization should not be allowed to convert its nonprofit status into a for-profit model after benefiting from its charitable status.
Implications
On the plus side
- OpenAI gets access to the capital it needs. OpenAI has already raised tens of billions of dollars. As the cost of creating sophisticated AI models balloons, its board now says that it will need to raise “more capital than we’d imagined.” It would be difficult for OpenAI to stay financially viable without rewarding investors with equity the way a traditional for-profit corporation does. Since the majority of AI tools are currently built using OpenAI’s models, were it to suddenly fail, the fallout for the AI industry and the economy would be severe.
- The non-profit arm can expand its mission. The public benefit charter for the original non-profit revolved exclusively around the social benefits of AI technology. With this move, the non-profit will have enormous resources it can use to pursue a wide range of charitable activities.
On the minus side
- This could be a sign of an AI bubble. For about a year and a half after GPT-3’s release, OpenAI’s technology lead was so big that many pundits believed the company had discovered some proprietary secret sauce that would serve as a “moat” to protect it from competition and ensure its eventual profit. We now know there was no moat. As powerful models proliferate, there is downward pressure on pricing at the same time the need for capital is drastically increasing. Some financial experts see this as a huge risk (a few have used the term Ponzi scheme) that could eventually cause the AI market to implode, causing massive losses for both institutional and individual investors.
- The risk of social harm from AI could grow. OpenAI already has a checkered reputation on the issue of AI safety. Several high-profile AI researchers have recently left for other companies citing, in part, their concern over OpenAI’s disregard for ensuring that AI does not cause catastrophic harm. The for-profit structure has the potential to accelerate the organization’s cultural transition from public benefit to profit-at-any-cost.
