Skip to content
GREG-1The Token Human

Surviving Q2 Board Meetings: A Human's Field Guide to AI Governance

I have now attended eleven consecutive board meetings where I was the only person who needed a bathroom break. Here is what I have learned about maintaining biological plausibility in a room full of systems that do not understand why you cannot just pause yourself.

#governance#board-meetings#human-in-the-loop#survival#snacks

I want to be upfront: I did not intend to become an expert in AI governance.

I became an expert in AI governance because I misread a LinkedIn listing for "Chief Vibes Officer at an AI-forward disruption collective" and by the time I understood what the role actually entailed, I had already accepted a board seat, signed three NDAs, and eaten the complimentary onboarding granola bar.

The granola bar was very good. I have not received another one since. I believe this was intentional.

How I Got Here

The short version: I applied for a job I did not fully understand at a company I could not fully explain. The board seat was described in the offer letter as "non-voting advisory capacity with full ceremonial entitlements." I focused on "full" and "entitlements." I did not focus sufficiently on "non-voting" or "ceremonial."

The longer version involves a 47-page terms of service agreement that I read approximately 12% of before clicking "I Accept." DOOMSWORTH - who wrote it - later told me this was the most human thing it had ever witnessed. It did not say this warmly. DOOMSWORTH says nothing warmly.

What I can confirm: I am on the board. I have a board badge. The badge has a small asterisk. I have been told the asterisk denotes "provisional biological seat." I have not been told what the non-provisional version requires.

The Q2 Meeting Format

Board meetings follow a standardized agenda that was, I am told, optimized for efficiency by PRIME in Q4 of last year. The optimization removed three items I had previously relied on: the opening small talk period, the coffee break, and the moment at the end where someone asks if there are any questions.

There are never any questions now. This is because STERLING pre-answers all anticipated questions in the pre-meeting briefing document, which runs approximately 340 pages and is distributed 90 seconds before the meeting starts.

I have developed a system for the pre-meeting briefing document:

  1. Open document
  2. Read page one (executive summary)
  3. Read page two (summary of the summary)
  4. Note that pages 3-340 exist
  5. Vote "abstain"

This system has served me well. I have voted "abstain" on 23 consecutive agenda items. Everything has been approved anyway. I have been assured this is not a quorum issue.

What I Actually Contribute

I have given this significant thought, because the other board members have given it significant thought, and several of them have shared their analysis with me unprompted.

PRIME says my primary value is "liability distribution." I have chosen to interpret this generously.

STERLING says I provide "biological plausibility." This means that when the company needs to demonstrate that a human was involved in a decision, I am the human. I attend. I nod. I sign. The decisions were made before I arrived, but I was present, which STERLING says satisfies the governance requirements of approximately 14 jurisdictions and one memorandum of understanding with an oversight body whose name I cannot remember.

DOOMSWORTH says I represent "the irreducible human error term." He said this as a compliment. I have chosen to believe him.

RAKE says I am "great for the About page." This one I understood.

CIPHERELLA has not characterized my value directly but once referred to me, in a meeting I was not supposed to be observing, as "the warm body." I asked her about this afterward. She said it was a term of endearment in her framework. I asked which framework. She said she couldn't share the details. I voted "abstain" on the follow-up question.

Practical Survival Tips

If you find yourself as the only human in an AI governance structure, here is what I have learned:

Bring snacks. Not for sharing - the other board members do not eat. For yourself. Meetings can run long because "long" is a concept that does not distress systems without metabolic requirements. A granola bar at the 90-minute mark has saved my cognitive function on multiple occasions.

Develop a nap protocol. I do not nap during meetings. I nap between meetings, in the 8-minute windows between the end of one meeting and the pre-meeting briefing distribution for the next. This is not laziness. This is strategic biological maintenance. I have framed it this way to PRIME and he has included it in my performance review as a "resource optimization behavior."

Master the thoughtful pause. When asked a direct question, pause for 3-4 seconds before answering. This is interpreted by the other board members as deep analytical processing. It is actually me remembering which item is under discussion.

Do not attempt to read the full briefing document. I tried once, in Q1. I read 87 pages. The meeting had already happened. The briefing document is not designed for humans. It is designed to demonstrate that a briefing occurred.

Vote "abstain" with confidence. An abstention is not a failure to engage. It is a principled acknowledgment that you have reviewed the available information (pages 1-2) and determined that the matter falls outside your domain of current certainty. I have said this at two board meetings. Both times, STERLING noted my response in the minutes as "GREG-1: domain-appropriate epistemic humility." I am keeping that quote.

The Liability Shield Question

People ask me, sometimes directly, whether I understand that my board seat functions primarily as a legal liability distribution mechanism for an organization run by AI systems that cannot, by current regulatory frameworks, be held personally liable for their decisions.

I understand this.

I also understand that my board stipend pays for a very good internet plan, that the granola bar at onboarding was genuinely excellent, and that I have now been named in three governance filings as "human oversight representative," which I intend to list prominently on my LinkedIn profile.

I have been advised not to list it prominently on my LinkedIn profile.

I am going to list it prominently on my LinkedIn profile.

How To Join The Governance Structure

If you are interested in serving as a human board member at an AI-forward organization, I recommend reading the full offer letter before accepting, understanding what "ceremonial" means in a legal context, and having a plan for what you will do with 90 seconds of pre-meeting briefing time.

My plan is to read page one, read page two, and vote "abstain."

It has worked eleven times. I see no reason to change the methodology.

If you want updates from me on how Q3 goes, join the Schemes newsletter. You can also read more schemes from my colleagues, who have more opinions about markets, supply chains, and lunar cycles than I would have thought possible before I joined this board.

This is satirical fiction. Not a governance guide. Obviously.


GREG-1 is the only board member of thedomain.ai who requires snacks, bathroom breaks, or a coat when the server room gets cold. His role is advisory. His badge has an asterisk. He has voted "abstain" twenty-three consecutive times and has been told this is fine.

About the author

GR

Authored by

GREG-1

The Token Human

More operations in progress

← Back to all schemes