OpenAI is a nonprofit-business hybrid: A management expert explains how this model works – and the way it fueled the uproar surrounding the short-lived ouster of CEO Sam Altman

The board of OpenAI, the developer of the favored artificial intelligence tools ChatGPT and DALL-E, fired its CEO Sam Altman at the top of November 2023.

Chaos ensued as investors and employees rebelled. When the chaos subsided five days later, Altman had returned triumphant to the OpenAI group amid the euphoria of employees and three of the board members who had asked for his ouster had resigned.

The structure of the board — a nonprofit board overseeing a for-profit subsidiary — appears to have played a job within the drama.

As a research management scientist organizational responsibility, governance and performanceI would love to elucidate how this hybrid approach is imagined to work.

Hybrid governance

Altman is co-founder of OpenAI in 2015 as tax-exempt nonprofit organization with a mission “To construct artificial general intelligence (AGI) that’s protected and advantages all of humanity.” To raise more capital than it could raise through charitable donations, OpenAI later created a holding company that might allow it to boost money from investors for a for-profit subsidiary he founded.

OpenAI executives decided to do that “Hybrid governance“Structure that permits it to stay true to its social mission while harnessing the facility of markets to extend its operations and sales. By combining profit and purpose, OpenAI was in a position to raise billions from investors in search of financial returns while “balancing”Commerciality with safety and sustainabilityreasonably than specializing in pure profit maximization,” it said in an announcement on its website.

Large investors subsequently have a big share within the success of their business activities. This is especially true for Microsoft, which has since owned 49% of OpenAI’s for-profit subsidiary invested $13 billion in the corporate. But these investors usually are not entitled to board seats like they might in typical corporations.

And the profits that OpenAI returns to its investors are capped at roughly 100 times what the unique investors invested. This structure requires it return to a nonprofit organization once this point is reached. At least in principle, this design was intended to stop the corporate from straying from its goal of safely helping humanity and never to jeopardize its mission through the reckless pursuit of profit.

Other hybrid governance models

There are more Hybrid governance models than you may think.

For example the Philadelphia Investigatorsa for-profit newspaper, owned by the Lenfest Institute, a non-profit organization. The structure allows the newspaper to draw investment without compromising its purpose – journalism that serves the needs of local communities.

Patagonia, a designer and supplier of outside clothing and equipment, is one other distinguished example. Its founder, Yvon Chouinard, and his heirs will remain in perpetuity transferred their property to a charitable foundation. Patagonia’s entire profits now fund environmental protection projects.

Anthropic, one in every of OpenAI’s competitors, also has a hybrid governance structure, nevertheless it is structured in a different way than OpenAI’s. There are two different governing bodies: a company board and what it calls a long-term profit trust. Because anthropic is a Non-profit corporationits board of directors may also have in mind the interests of other stakeholders along with the owners – including the general public.

And BRACa world development organization founded in Bangladesh in 1972 one in every of the world’s largest NGOs, controls several for-profit social enterprises that profit the poor. BRAC’s model is comparable to OpenAI’s in that a nonprofit organization owns for-profit corporations.

Origin of the board’s conflict with Altman

The nonprofit board’s primary responsibility is to make sure that the mission of the organization it oversees is maintained. In hybrid governance models, the board must make sure that market pressures to make cash for investors and shareholders don’t override the organization’s mission – a Risk often known as mission drift.

Nonprofit boards have three primary responsibilities: the duty of obedience, which requires them to act within the interests of the organization’s mission; the duty of care, which requires them to exercise due care when making decisions; and the duty of loyalty, which requires them to avoid or address conflicts of interest.

It seems that the board of OpenAI tried to exercise the duty of obedience when it was decided to fireplace Altman. The official reason given was that he “He was not at all times open-hearted in his communications” together with his board. Additional justifications provided anonymously by people identified as “Concerned former OpenAI employees” weren’t checked.

Additionally, Board member Helen Tonerwho resigned from the board within the midst of those upheavals, Co-author of a research paper only a month before the failed try and depose Altman. Toner and her co-authors praised Anthropic’s precautions and criticized OpenAI’s “frantic austerity measures” around the discharge of its popular chatbot ChatGPT.

Mission vs. money

This was not the primary try and oust Altman on the grounds that he was straying from his mission.

In 2021, the organization’s head of AI security, Dario Amodei, tried to persuade the board of directors to accomplish that Push Altman out for security reasons, shortly after Microsoft invested $1 billion in the corporate. Amodei later left OpenAI together with a few dozen other researchers and founded Anthropic.

The push and pull between mission and money is probably best embodied by Ilya Sutskever, a co-founder of OpenAI, its chief scientist and one in every of three board members who were forced out or resigned.

Sutskever initially defended the choice ousting Altman on the grounds that this was mandatory to guard the mission of constructing AI useful to humanity. But he later Changed his mindHe tweeted: “I deeply regret my involvement within the board’s actions.”

He eventually signed the worker letter demanding Altman’s reinstatement and stays the corporate’s chief scientist.

AI risk

An equally vital query is whether or not the board has fulfilled its duty of care.

I believe it’s reasonable for OpenAI’s board to query this The company has released ChatGPT with sufficient guardrails in November 2022. Since then, large language models have caused chaos in lots of industries.

I saw this primary hand as a professor.

In many cases, it is nearly not possible to detect whether students are cheating on assignments through the usage of AI. Admittedly, this risk pales as compared to AI’s ability to do even worse things, like help with design Pathogens with pandemic potential or create Disinformation and deepfakes that undermine social trust and endanger democracy.

On the opposite hand, AI has enormous potential Benefits for humanityfor instance, accelerating the event of life-saving vaccines.

But the potential risks are catastrophic. And once this powerful technology is released, nothing more will probably be known.Off switch.”

Conflicts of interest

The third duty, loyalty, is dependent upon whether board members have conflicts of interest.

The most blatant was that they desired to make cash from OpenAI’s products, so could they jeopardize its mission within the expectation of monetary gain? Typically the members are one Nonprofit board members are unpaid, and people who don’t work for the organization don’t have any financial stake in it. CEOs report back to their board of directors, which has the authority to rent and fire them.

Until the recent restructuring of OpenAI Three of the six board members were paid executives – the CEO, the chief scientist and the president of its for-profit arm.

It doesn’t surprise me that while the three independent board members all voted to remove Altman, all the paid executives ultimately supported him. Earning your paycheck from an organization you might be imagined to oversee is taken into account Conflict of interest within the non-profit world.

I also imagine that even when OpenAI’s reconfigured board succeeds in fulfilling the mission of serving the needs of society reasonably than maximizing its profits, it will not be enough.

The technology industry is dominated by corporations like Microsoft, Meta and Alphabet – large, for-profit corporations, not mission-driven nonprofits. Given the risks, I imagine effective regulation is required – leaving management to AI developers won’t solve the issue.