OpenAI's Rollercoaster Ride and the Root Cause

Sam Altman is back at the helm of OpenAI following a tumultuous week marked by internal strife and power struggles. The board and internal politics seem to have attempted a coup d'état, largely fueled by concerns that he was moving too quickly with product development and new AI features. They feared that the rapid introduction of new AI features and products might spiral out of control for OpenAI and possibly the world.

In terms of the timeline, it was swift. On a Friday, they fired Sam Altman. There was almost no warning. Even Microsoft, which had invested 10 billion USD, was reportedly only given scant notice—a mere 15 minutes. Sam temporarily joined Microsoft’s AI team, while hundreds of OpenAI staff made it clear they were ready to leave if Sam was not reinstated at OpenAI. In the end, it was the board, or at least three of its members, who were really the ones getting fired. Ultimately, this was a reversal of the board's decision, and in an ironic twist, it was the board members who faced dismissal instead of Altman. The next consequence will now be a public discourse that Sam Altman and OpenAI will be unchecked and accelerating even faster towards a potentially harmful version of AI. And of course, many will disagree that any real danger is imminent.

Regardless of the new discourse, what is the root cause of this entire episode? One hypothesis is that OpenAI is really a company aimed at making a profit and exists within a non-profit foundation. The board members have incentives and directions that are completely opposite to those of the OpenAI commercial enterprise. The majority of the board is independent, and the independent directors do not hold equity in OpenAI. The board’s responsibility is to advance OpenAI's mission and preserve the principles of its Charter, which is publicly communicated, "to ensure that artificial general intelligence—AI systems that are generally smarter than humans—benefits all of humanity."

At the same time, investors like Microsoft, Sequoia, and others are not likely to be looking for their AI investment to slow down to a crawl while competitors like Google and others might race ahead. Certainly some politics has been involved, but there has also been a collision of the rules and motivations of the private enterprise and the non-profit. And in this case, it is the non-profit oriented board members who seems to be the main casualty.

You might ask what this means about AI and about whether companies can and should self regulate. ?Even if they tried ?to take a balanced pace of AI development, other firms would race ahead of them. ??In general, hard for any organization to be both a regulator and the regulated. ?It’s ok for a natural tension to exist between the two ?

Further, OpenAI is a special case where they act as ?both a for profit and non profit?organization, which is rare and again, it’s hard for organizations to be both.? So far that has not been successful in the past.?

OpenAI is now going through the conversion process towards a largely commercial enterprise.? Or at least one where the non profit organization will have more separation from the commercial agenda, which is the case with most other private foundations today.? Perhaps the open part of OpenAI should actually be open and supply knowledge and resources to everyone including competitors, while enabling the commercial activity to remain proprietary and competitive, and while allowing regulation by the same actual regulators who regulate everyone else. ?

Rosa Aranda Barrio Rafif Srour Guillermo de Haro Rodríguez Geoffroy Gérard Pamela Rolfe