#13: A Timeline of the OpenAI Board

Yesterday, Sam Altman and Greg Brockman were fired from the Board of Directors of OpenAI. Following, all of Tech Twitter was abuzz with one question: wait a moment, who was on the Board? And after they found out, they asked: who on earth are Tasha McCauley and Helen Toner? It turns out that OpenAI’s Board had undergone numerous changes over the years, especially recently. And that just wasn’t ever the biggest news about OpenAI, so those changes didn’t spark the concerns that maybe they should have.

I combed through the Internet Archive and OpenAI’s non-profit filings to try to make sense of OpenAI’s governance. Below, I have attempted to chronicle the composition of OpenAI’s Board over time, point out the conflicts, and you can see how we got to the earthquake yesterday. You can skip to the end for my summary perspective.

December 11, 2015

OpenAI is founded.

Board Directors:¹

• Elon Musk (Co-Chair)

• Sam Altman (Co-Chair)

December 31, 2016

OpenAI’s Form 990 public filings for calendar year 2016 show the Board Directors:

• Elon Musk

• Sam Altman

• Chris Clark

• Jonathan Levy (?)

Chris was the initial COO of OpenAI, and still works there to this day. Jonathan Levy was listed as Secretary/Treasurer, and may have been a trustee rather than a Director. It’s unclear from the filings.

March 2017

Open Philanthropy donates $30M to OpenAI. Holden Karnofsky, the founder of Open Philanthropy, joins OpenAI’s Board of Directors.

December 31, 2017

OpenAI’s Form 990 public filings for calendar year 2017 show the Board Directors:

• Elon Musk

• Sam Altman

• Chris Clark

• Holden Karnofsky

• Greg Brockman

• Ilya Sutskever

February 20, 2018

Elon Musk is removed from the Board. The official press release proclaims a departure to avoid potential conflicts, but journalists report leadership disagreements culminating in Elon proposing a takeover and being rebuked.

Board Directors:

• Greg Brockman

• Ilya Sutskever

• Holden Karnofsky

• Sam Altman

We don’t know exactly when Chris Clark was removed from the Board.

March 2018

Reid Hoffman, founder of LinkedIn and General Partner at Greylock, joins the Board. I couldn’t find a press release or official announcement, but Reid’s LinkedIn profile has the dates.

April 24, 2018

Adam D’Angelo, CEO of Quora and former Facebook CTO, joins the Board. This follows the February Board changes, where the OpenAI blog post had noted the intent to add another Director to the Board soon.

September 2018

Sue Yoon joins the Board. Sue’s exact employment at the time was unclear — she was previously an EIR at First Round, and in the coming months would lead robotics projects at Google. Similar to Reid Hoffman, I couldn’t find an official announcement, but her LinkedIn profile has the dates.

December 31, 2018

OpenAI’s Form 990 public filings for calendar year 2018 list the Board Directors:

• Sam Altman

• Sue Yoon

• Holden Karnofsky

• Greg Brockman

• Ilya Sutskever

• Adam D’Angelo

• Tasha McCauley

I could not find anything in the way of a source on when, or under what circumstances, Tasha McCauley joined the Board.

March 11, 2019

This gets strange. There’s an OpenAI blog post listing the Board Directors:

• Greg Brockman

• Ilya Sutskever

• Sam Altman

• Adam D’Angelo

• Holden Karnofsky

• Reid Hoffman

• Shivon Zilis

• Tasha McCauley

Note the unannounced elevation of Shivon Zilis (previously an advisor) and the unannounced departure of Sue Yoon. Weirder yet, OpenAI published its new homepage just that day, still listing Sue Yoon as a Board Director, and not Shivon Zilis.

November 2019

Sue Yoon leaves OpenAI’s Board, according to her LinkedIn. The OpenAI Website still lists her (and not Shivon Zilis) as a Board Director.

December 31, 2019

Another year, another OpenAI Form 990 public filing, listing the Board Directors:

• Ilya Sutskever

• Greg Brockman

• Sam Altman

• Reid Hoffman

• Sue Yoon

• Holden Karnofsky

• Adam D’Angelo

• Tasha McCauley

Note that Shivon Zilis still doesn’t appear in the list of Board Directors. Was the March 11, 2019 blog post just wrong? Did someone in marketing make a mistake and no-one caught it?

December 31, 2020

Slow news year. The Form 990 public filing for calendar year 2020 lists the Board of Directors, finally including Shivon Zilis and not Sue Yoon:

• Ilya Sutskever

• Greg Brockman

• Sam Altman

• Reid Hoffman

• Shivon Zilis

• Holden Karnofsky

• Adam D’Angelo

• Tasha McCauley

I couldn’t find a public statement on when Shivon actually joined the Board, other than the March 2019 blog post that may have been in error.

May 3, 2021

Will Hurd, Republican member of the House of Representatives, and former CIA agent, joins the Board.

September 8, 2021

Helen Toner, Director at Georgetown’s Center for Security and Emerging Technologies, and formerly of Holden Karnofsky’s Open Philanthropy, joins the Board.

Fall 2021

Holden Karnofsky resigns from the Board, citing a potential conflict because his wife, Daniela Amodei, is helping start Anthropic, a major OpenAI competitor, with her brother Dario Amodei. (They all live(d) together.) The exact date of Holden’s resignation is unknown; there was no contemporaneous press release.

Between October and November 2021, Holden was quietly removed from the list of Board Directors on the OpenAI website, and Helen was added (Discussion Source). Given their connection via Open Philanthropy and the fact that Holden’s Board seat appeared to be permanent, it seems that Helen was picked by Holden to take his seat.

December 31, 2021

OpenAI’s Form 990 public filings list the Board Directors of the 2021 calendar year:

• Ilya Sutskever

• Shivon Zilis

• Greg Brockman

• Will Hurd

• Sam Altman

• Reid Hoffman

• Holden Karnofsky

• Adam D’Angelo

• Tasha McCauley

• Helen Toner

The fact that both Holden and Helen are listed here is not surprising; both of them were Board Directors at points in 2021. (It does not necessarily imply that they were both on the Board at the same time.)

2022?

There did not appear to be any Board events in 2022. The Form 990 does not appear to have been filed as of the time of writing.

January 2023

Reid Hoffman steps down from the Board, citing the need to avoid potential conflicts with his investments. While this was reported in March 2023, according to his LinkedIn profile’s dates it happened in January.

March 23, 2023

Shivon Zilis resigns from the Board for reasons unknown. (Commentators speculate that her resignation is over conflicts due to her bearing Elon Musk’s children, but that is ultimately just speculation.)

July 13, 2023

Will Hurd resigns from the Board, citing the need to focus on politics/his 2024 Presidential campaign. (Three months later, in October, he drops out of the race. I don’t know what to make of that.)

November 17, 2023

Sam Altman is fired from OpenAI and the OpenAI Board in a surprise meeting of the Board (except Greg). Minutes later, in a separate surprise Board meeting, Greg Brockman is removed from the Board (and as Board Chairman).

Board Directors:

• Adam D’Angelo

• Helen Toner

• Tasha McCauley

• Ilya Sutskever

Summary/Perspectives

The first thing that sticks out to me is that there have been, for several quarters, two significant conflicts of interest on the Board:

• Adam D’Angelo founded and appears to be spending all his time on developing Poe, an AI chat platform partially leveraging and partially competing with OpenAI. In my opinion, that’s too close. Reid Hoffman resigned over potential indirect investment conflicts; Adam’s conflicts are more direct. Best practice would’ve been for Adam to resign when he began working on Poe.

• Helen Toner and and Tasha McCauley are jointly participating in a highly ideological AI governance organization. As Alex Konrad noted: “McCauley currently sits on the advisory board of British-founded international Center for the Governance of AI (GovAI) alongside fellow OpenAI director Helen Toner.” It turns out that the advisory board is six people, and beyond Helen and Tasha, the other four include: one who currently works for Open Philanthropy, and another is the founder of GovAI, which was mostly funded by… Open Philanthropy.

  • For OpenAI’s six-person Board, it was inappropriate for two Board Directors to be this strongly associated with an ideological organization and therefore so strongly and predictably aligned in their voting. It calls into question the independence of their votes.

  • Due to Open Philanthropy’s link to major OpenAI competitor Anthropic, there’s also a hint of corporate conflict here. If I were on OpenAI’s Board, I would have requested at least for Tasha² to relinquish her seat for a true independent Director.

Secondly, what the hell happened in Q1/Q2 2023?

• Reid, Shivon, and Will all resigned, and the Board did not line up replacement Directors? By comparison, when Elon resigned in February 2018, Adam joined two months later.

  • Were these seats just left vacant, with a deadlocked Board unable to agree on new Directors to appoint?

• They all resigned within a few months of one another despite OpenAI looking like the rocketship of the century? Something feels a little odd about that.

It seems less likely that the November firings would have happened if Reid, Shivon, and Will — or even just one of them! — had still been on the Board, or replaced with an appropriate representative. With this view, the outcome was almost predictable given these two facts:

1. The thinning-out of the Board from 9 to 6 members;

2. Half of those 6 members carrying conflicts in their relationship with OpenAI!

We will find out, in due time, the motivations of the Board in the November firings. Right now they aren’t clear. It isn’t known whether anyone acted inappropriately, and I am not accusing anyone (to be clear, even the Board Directors that I consider conflicted) of having acted subject to conflicts of interest. But the 2023 changes made drama likely, no matter what. A Board is a delicate balance of perspectives and interests. When a Board rapidly changes in size, rarely is the remainder left well-balanced. Potential conflicts only make the balancing act harder.

Final Thought

Governance can be messy. Time will be the judge of whether this act of governance was wise or not. But you should note that the people involved in this act of corporate governance are roughly the same people trying to position themselves to govern policy on artificial intelligence.

It seems much easier to govern a single-digit number of highly capable people than to “govern” artificial superintelligence. If it turns out that this act of governance was unwise, then it calls into serious question the ability of these people and their organizations (Georgetown’s CSET, Open Philanthropy, etc.) to conduct governance in general, especially of the most impactful technology of the hundred years to come. Many people are saying we need more governance: maybe it turns out we need less.

1. Source is this Vanity Fair interview with Sam Altman, where he notes that the only Directors are himself and Elon: https://www.vanityfair.com/news/2015/12/sam-altman-elon-musk-openai

The official blog post confirms both Elon and Sam as co-chairs: https://openai.com/blog/introducing-openai

2. Ideally I would have requested it of both Helen and Tasha, but it seems that Helen’s seat was bought-and-paid-for, so that might not have gotten a lot of traction. However, it appears that Tasha was actually meant to be an independent Director!