Bilateral AI Accord Frontier Explorer

Analyze diplomatic bargaining feasibility, compute the Zone of Possible Agreement (ZOPA), and test treaty resilience across lethal autonomy, CBRN safeguards, and frontier compute monitoring.

ZOPA Feasibility
Narrow
Overlap area: 18.4%
US Net Utility
+14.2
Exceeds reserve (+5.0)
China Net Utility
+9.8
Exceeds reserve (+5.0)
Defection Fragility
High (62%)
Vulnerable to covert training
Bilateral Bargaining Frontier (Nash Equilibrium)
US Reserve Line
China Reserve Line
ZOPA Consensus Zone
Current Accord Position
Bilateral consensus is mathematically viable but constrained to low-inspection, existential risk boundaries. Nash Optimum: US 64.2 pts | China 58.6 pts
Treaty Clause US Net Stance China Net Stance Verification Friction Consensus Verdict
Interactive model computed. 6 clauses evaluated across strategic parameters.

Why AI Consensus Narrowed

As highlighted in recent bilateral diplomatic reporting, dialogue between the United States and China on artificial intelligence is caught between existential containment and technological sovereignty. Finger-pointing over advanced compute export curbs and cyber operations reduces diplomatic space.

This explorer adapts classic Rubinstein-Nash non-zero-sum bargaining models to frontier AI dilemmas: evaluating when shared aversion to accidental catastrophe outweighs the fear of giving up military asymmetry.

Model Levers & Verification Dynamics

Zone of Possible Agreement (ZOPA)

ZOPA emerges only when the cumulative utility of an accord exceeds both nations' best alternatives to a negotiated agreement (BATNA / reservation thresholds). Verification distrust shrinks this overlap dramatically.

Asymmetric Compute & Sovereignty Concerns

Frontier cluster on-site inspections trigger deep espionage concerns for Beijing, while unverified self-reporting triggers defection fears for Washington. Accords focused on redlines (CBRN and lethal autonomy) preserve higher feasibility than cluster quotas.

Crisis Communication Hotlines as Anchor Agreements

When high distrust scuppers intrusive technical audits, mutual hotlines and unclassified incident reporting mechanisms remain resilient pareto-optimal starting points.

Enjoy this tool? Build your own with Super