Who holds it
75 · Any group
Walk Part III: law, firms, states, alignment, long-term risk. Power is not a later chapter. It is already in the room.
Opening question
If this goes well, who is in a position to say so? If it goes badly, who can be named?
To hold together
- What is the difference between a rule we can contest and a terms-of-service we can only click?
- When a handful of firms hold the stack, what is left for a town, a school, or a clinic to decide?
- How do we speak about distant, severe possibilities without letting them eat the near ones?
A practice
Open Part III. Read a governance concern and a long-term concern. Place them on a table. Ask which one our group is actually in a position to act on this month.
Concerns to name in the room
Youth and future-generations representation
AI governance can overlook young people and future generations when decisions prioritize current adult interests and short-term results.
Uncertainty and disagreement about probability
Disagreement about catastrophic AI risk reflects uncertainty about future capabilities, human choices and the ways events could unfold.
Principles to sit beside the facts
Accountability
Name who holds the consequence before it is needed.
Stewardship
Hold what is built in trust for those who will live with it.
Humility
The map is larger than any one of us.
Courage
Act without the comfort of certainty.
Read or watch beforehand
One piece is enough. Save the rest for later.
Alondra Nelson
Former deputy director, White House OSTP · Institute for Advanced Study
Shepherded the Blueprint for an AI Bill of Rights. A civic starting point for rights language without requiring a legal training.
Photo credit
Executive Office of the President of the United States · Public domain (US government). Displayed in a circular crop; the photo retains its stated license.
International Committee of the Red Cross
Humanitarian policy on AI
Writes from international humanitarian law, not from product. Essential when a conversation touches conflict, weapons, or civilian protection.
Stuart Russell
Professor of computer science · UC Berkeley · CHAI
Argues that the standard model of AI, optimizing a fixed objective, is the wrong shape, and proposes machines that remain uncertain about human preference.
Photo credit
Bengt Oberger · CC BY-SA 4.0. Displayed in a circular crop; the photo retains its stated license.
NIST AI Risk Management Framework
Voluntary framework · U.S. National Institute of Standards and Technology
The document many institutions already have to map against. Not a voice in the personal sense, but the shared language of governance conversations.