LiveHuman-AI & attentionWestern cycleAnchors pending

International AI-governance treaty analogous to NPT proposed by ≥3 major powers by 2028, though not ratified by 2030?

As of 6 Oct · 22:12 UTCResolves 31 Dec 20282 movers0 pressure moments
Timeline A · Governed AIReality
60%
Crowd 20%Gap −40

By late 2028, at least three major powers—likely the U.S., EU, and China—have put forward a formal treaty proposal that mirrors the Nuclear Non-Proliferation Treaty's structure: commitments to AI transparency, restrictions on dual-use model development, verification mechanisms, and tiered obligations for leading vs. developing nations. The proposal exists as a signed diplomatic text, has been tabled at the UN, and is debated in national legislatures, but ratification votes have not yet occurred or have stalled. The Bletchley and Seoul declarations evolved into binding legal language.

Timeline B · AI RaceReality
40%
Crowd 80%Gap +40

No treaty proposal matching NPT-style design emerges from major powers by the end of 2028. Instead, governance remains fragmented: bilateral agreements, regional frameworks (EU AI Act enforcement, UK-Singapore partnerships), industry codes of conduct, and non-binding UN resolutions continue to fill the space. China, the U.S., and the EU pursue separate strategic AI standards rather than converging on a single multilateral treaty model. The diplomatic momentum from Bletchley and Seoul fades into advisory reports without treaty architecture.

crowd 20
The quiet paththe one nobody’s watching

A state-sponsored AI attack shatters treaty negotiations and forces military-first protocols instead.

6%
The read
Reality leans toward governed ai.

Reality points at Governed AI — the dominant pressures today support that outcome over AI Race. The crowd lands lower on Governed AI than the evidence supports.

Where reality is heading

A, 80% bandCrowd today
Projected reality support to 31 Dec 2028A rises from 41% to about 52% with an 80% band of 33 to 71. B falls from 53% to about 44%. Quiet path near 4 to 6%. Crowd today: A 55%.100%7550250Crowd today · A 20%A ~52%B ~44%P ~4%NowJan 2027Jan 202831 Dec 2028
Reality crosses toward A in early 2028. Today’s crowd number sits at the top of A’s band, possible but optimistic. View as table

The movers, ranked by leverage

01
AI safety-testing pre-deployment regulatory severityDetermines deployment-safety trade-off
leverage 0
Conditional
02
Chip export-control severityDetermines capability-race pace + China compute-substrate isolation
leverage 0
Conditional

What would change our mind

Signals toward A
  • attention-economy regulated (smartphone age-limits, algorithmic transparency, data-protection)
  • cognitive-dependency addressed via educational integration + digital-literacy
  • AI safety-testing pre-deployment regulatory severity: movement in the direction that supports Governed AI
Signals toward B
  • AI development treated as national-champion race
  • safety-frameworks minimized as constraint-on-innovation, sovereignty-first (US-AI, China-AI, EU-AI, potentially India-AI, Gulf-AI as distinct capability stacks), export-controls hardened, capability-race accelerated
  • AGI-timeline compressed as race dynamics dominate
  • attention-economy less regulated except where national-security implications
  • cognitive-dependency mostly market-mediated with cultural-restorationist opt-out

Evidence