AI governance cannot be left to industry alone; it requires an international, multi-stakeholder framework to prevent governance failure as AI capabilities advance.

Leaders from moral and industry circles, including voices from the Vatican and Anthropic, urge disarmament, regulation, and candor about AI risks, highlighting a shared urgency.

The overarching message is urgent: act now before catastrophe to protect humanity from AI-enabled risks.

Governments and international organizations should convene a governance process immediately, bringing in industry practitioners who understand failure modes to translate risk into binding commitments.

Current governance is internal to companies and lacks external enforceability, creating a collective action problem where no actor bears the full cost of responsible restraint.

A bilateral U.S.-China framework is essential but not sufficient; a broader coalition is needed to sustain global restraint and accountability.

Rather than a nuclear-arms analogy, AI governance should resemble a financial-system model like FATF, with shared norms, cross-border obligations, expert review, and real consequences for non-compliance.

With AI potentially reaching Artificial General Intelligence by around 2030, pre-emptive, collective governance is crucial to avert catastrophic outcomes.