The best way to stop an AI apocalypse might be to copy a 1,000-year-old tribal government system from a tiny Pacific island.
March 25, 2026
Original Paper
Governing AI in a Multipolar World: A Moral Parliament Grounded in Palauan Polycentric Governance
SSRN · 6193678
The Takeaway
As modern nations fail to coordinate on AI safety, this paper suggests the ancient "Kelulau" system—which uses a two-chamber parliament to balance high-tech developer autonomy with a collective stakeholder veto—is more effective at managing existential risk than current technocratic or international legal frameworks.
From the abstract
<p>AI governance faces a multipolar coordination crisis. Rapid capability advances outpace institutional response amid competing state and corporate interests unwilling to cede sovereignty. This paper proposes a moral parliament modeled on Palau's millennia-tested Kelulau system: a polycentric governance architecture balancing clan autonomy with collective action against catastrophe.</p> <p>The two-chamber design distributes power strategically: a Small Chamber of frontier AI-developing states/c