Leave-Behind · Policymakers · 2026
One-Minute Briefing
For staff who will not read a white paper. Hand this over. Speak the numbers. Sit down.
500+
Experts signed the 2023 CAIS statement ranking AI extinction risk with pandemics and nuclear war
$190M
Approx. yearly AI safety spend, against $250B+ into making systems more powerful
0
Binding global treaties on superintelligence development today
The ask
Three moves
- Domestic law: license large training runs; require independent pre-deployment evaluation.
- Multilateral talks: open negotiations on a prohibition with verification.
- Funding: expand independent evaluation and treaty-design capacity outside lab walls.
The frame
Prevention · narrow AI safe
We are not asking to freeze medicine or software. We ask to prohibit autonomous general systems past a capability line until control is solved. Precedents: Nuclear Non-Proliferation Treaty, chemical weapons ban, Montreal Protocol.
Anticipated objection: "China will not play along."
Then build verification (chip tracking, cluster inspection) and treat defection as an enforcement problem. A unilateral sprint to an uncontrollable system is not a strategy.
Staff briefing: nakadafoundation.org/contact/ · Plan: nakadafoundation.org/our-plan/ · Policy PDF: nakadafoundation.org/resources/pdf/policy-brief/