OpenAI is an open
M. Damodaran Chairperson, Excellence Enablers Former Chairman, SEBI, UTI and IDBI
01 Dec 2023
There are a number of Corporate Governance lessons that the OpenAI crisis provides us with.
Editor’s note: This was not written using ChatGPT.
It is often said that truth is stranger than fiction. The developments in the recent past at OpenAI is an open and shut case in support of this hypothesis.
There have been earlier instances in corporate history where high profile CEOs were removed from the company, and came back to occupy the same positions after several years had passed. This is the first case in living memory where a CEO was shown the door by the Board of Directors, offered a position in a large company with 49% shareholding in OpenAI, and came back within four days to hold the same position, in the process dismissing almost his entire Board of Directors. A stranger case of musical chairs is not easy to contemplate.
The stated reason for the removal of Sam Altman, the CEO, was that he was not “consistently candid” with the Board in regard to his various moves in the company, hindering its ability to discharge its responsibilities. The removal, which incidentally happened on a virtual call on a Friday, to which Greg Brockman, the Chairman of the company, does not seem to have been invited, and post which, he too was removed from his post, gave rise to more than 500 of the employees threatening to cast their lot with the expelled CEO, and in the process moving to Microsoft or wherever else he was going to be. By Tuesday, Sam Altman was back, and he promptly got rid of the Board that had expelled him, retaining only one Director in the reconstituted Board. Interestingly, the statement that Altman was being removed was read out by Mr Ilya Sutskever, who worked with Altman for 8 years.
Altman’s return to the company was not without its share of excitement. While pitching for the removal of the existing Board, Altman had demanded a Board seat for himself, which was not agreed to. His co-founder, who was earlier the Chairman, too lost his Board seat in the process of negotiation.
One stated reason, for the difference in regard to direction, between the earlier Board and the CEO, was that OpenAI, which was a not-for-profit organisation, had been brought into the for-profit space through a subsidiary. This had been the subject matter of discord within the Board. Some participants, as well as observers, of the ongoing drama expressed concern at the danger inherent in AI assuming a significantly larger presence in our lives. Reportedly, some employees were also concerned that the unbridled growth of AI would negatively impact human lives and security. Mira Murati, who was appointed interim CEO, when Altman was fired, was reportedly one of the persons who had earlier expressed some concerns to the Board.
While negotiations were on to bring Altman back, the Board named a new CEO, Emmett Shear. He could not step into the leadership role because Altman came back to his earlier role. Succession planning, anyone?
There are a number of Corporate Governance lessons that the OpenAI crisis provides us with, and it is through the lens of Corporate Governance that we propose to look at the various developments.
To begin with, as in the case of almost every Corporate Governance meltdown, notice has to be taken of the composition and the functioning of the Board. OpenAI had a 9-member Board. 3 of the Directors left the Board more than a year ago, and those positions were not filled. One of the Directors is stated to have expressed the view that he was uncomfortable with the way in which AI was sought to be used, and its possible negative influence on the world. The vacancies were not filled, and meanwhile, the CEO continued doing what he thought was right. For the Board to suddenly express concern in public, leading to abruptly expelling the CEO, should raise eyebrows. Information is the fuel on which corporate entities run, and if the management was not providing timely and complete information to the Board, action should have been taken much earlier than the time at which the CEO was removed. Even this removal seems to have been triggered inter alia by complaints by employees that the new initiatives were detrimental to public interest. It is reasonable to conclude that the Board was acting in an information vacuum that did no credit to it, and to the management. The related question, which cannot be ignored, is the representative role that the Board of Directors is expected to perform. Conventional Corporate Governance theory requires that the Board, acting on behalf of, and in the interest of, shareholders and other stakeholders, should manifest the ability and willingness to provide superintendence, direction and control. This was a role that the Board either did not perform, or could not perform, on account of information being denied to it, or being resisted by a charismatic CEO, demonstrating significant growth in the company. There is no escaping the fact that the Board had abdicated its responsibilities on a continuing basis. The demand that Altman should be replaced had been made by at least one Board member a year before the defining Friday call.
Let us now look at what Sam Altman did. Clearly, he did not keep his Board informed, leave alone enlightened. He seems to have not carried conviction with a portion of his workforce who felt that some of his moves were far too adventurous to be consistent with public interest. Was he then a lone ranger, with not many others persuaded to think along the lines that he had decided on? Was this a case of the strong CEO model that many persons in the USA swear by, carried to an extreme, where there was one power centre in the organisation, and nothing else? His co-founder, who was Chairman and President, seems to have backed him, but to no avail.
The fact that the decision to remove Sam Altman was taken seemingly abruptly, at a virtual meeting, to which the Chairman was not invited, gives rise to some other questions. Did members of the Board feel that they were inadequate to deal with a corporate leader of this kind? Was relative insecurity at the root of the extreme decision to sack him, rather than to erect a system of checks and balances to see that he did not run riot in the organisation? Clearly, the Board was divided in its assessment of Altman. The reconstituted Board retained one member, while getting rid of the rest. Were there conclusive conversations in the previous Board, with differing viewpoints, on the direction in which the company was headed? Not enough of this is in the public domain. It might be useful for regulatory agencies to carry out a deep dive, in order to determine what the problems were, so that similar events do not take place elsewhere in the corporate sector. Whether the new reconstituted Board will act independently remains to be seen.
In all of this what did Microsoft, with a 49% shareholding, and its celebrated CEO do? The Board did not seemingly keep the investors informed prior to sacking Altman. Having got the news, should Microsoft have welcomed Altman with open arms, rather than push back to have him retained as CEO? Was the largest shareholder, but one without a Board seat, oblivious of the interest of other investors and stakeholders? At least one other institutional investor seems to have expressed concerns a few months earlier. Should the institutional investors not have come together to tell the Board where it got off?
There is also the question of regulation as a check to unbridled adventurism. Off and on, in several jurisdictions, it has been argued by a section of the ecosystem, that external regulation is counterproductive, and that organisations should be left alone to regulate themselves. This stance has been adopted by persons who ought to be better informed, having regard to number of instances where self-regulation, especially when it conflicts with business interests, has been found to be wholly inadequate. In the context of the explosive growth of AI and its various applications, and the impact that it could have on human life and work, there is a need to put in place effective regulations by Government, especially having regard to the fact that there are security aspects that need to be addressed. Effective Corporate Governance cannot be predicated on self-regulation. Nothing seems to have happened by way of regulatory intervention, proving yet again that the policeman is the last person to reach the scene of crime.
There are perhaps some winners and some losers in this saga. What is abundantly clear is that Corporate Governance has taken a bad knock.