Montreal Protocol
The Montreal Protocol is the 1987 international treaty that phased out the production of chlorofluorocarbons and other ozone-depleting substances, substantially reversing the damage to the stratospheric ozone layer that had threatened human and ecological health. The protocol is the most successful international environmental treaty in history, universally ratified and substantially achieving its objectives. For Gore, it is the foundational precedent establishing that democratic societies can successfully govern transformative technology risks when the scientific evidence is clear, technological alternatives are available, and political will is mobilized. The protocol's success is the empirical refutation of the claim that international governance of powerful technologies is impossible.
The Protocol's Conditions Are Absent — Contrarian ^ Opus
There is a parallel reading of the Montreal Protocol — as a testament not to what democratic coordination can achieve, but to the exceptional conditions under which it becomes possible. The protocol succeeded because CFCs were a narrow technical input to a limited industrial sector, produced by a small number of manufacturers who could be brought to the table. The substitute technologies were owned by the same companies that manufactured CFCs — DuPont's shift to HFCs was not self-sacrifice but market repositioning. The economic stakes were manageable; the geopolitical stakes minimal. No major power derived its strategic position from CFC production. The protocol's success depended on a technical problem that could be solved without threatening core economic or political arrangements.
AI presents the inverse configuration at every point. The technology is general-purpose, integrated across every sector, developed by thousands of entities in dozens of countries. No "alternative" exists because AI is not an input to some other process — it is becoming the substrate of economic activity itself. The economic stakes are not manageable but civilizational; the geopolitical stakes are the primary axis of great power competition. The United States and China are not negotiating AI governance in good faith — they are racing for technological dominance while performing concern about safety. The Montreal Protocol is not a precedent for AI governance. It is a demonstration of how far the conditions for AI governance are from anything history has produced.
In the AI Story
The protocol's operational features distinguish it from subsequent climate efforts. Clear empirical triggers — measured ozone levels — connected scientific evidence to regulatory response without requiring renewed political negotiation. Industry alternatives — hydrofluorocarbons and other substitutes — were available when the protocol took effect, making compliance economically feasible. Enforcement mechanisms — trade sanctions against non-complying nations — created real consequences for defection. Scientific infrastructure — the World Meteorological Organization's ozone monitoring network — provided the independent verification that distinguished real compliance from reported compliance.
Gore invokes the protocol repeatedly as the operational template for what AI governance could achieve. Empirical triggers: specific capability thresholds or deployment conditions that automatically invoke regulatory response. Available alternatives: safety research, alignment techniques, and deployment restrictions that are technically feasible when required. Enforcement mechanisms: real consequences for companies or countries that defect from agreed frameworks. Scientific infrastructure: the kind of independent monitoring that Climate TRACE represents for emissions, applied to AI deployment patterns and effects.
The protocol's relationship to the Paris Agreement is instructive. Montreal addressed a narrower problem with clearer alternatives, enabling more ambitious commitments and more effective enforcement. Paris addressed a vastly more complex problem with more politically difficult alternatives, producing weaker commitments and enforcement. The difference is not that Paris failed — it has produced real effects — but that the more ambitious Montreal model required conditions that the climate problem did not permit. AI governance probably falls somewhere between the two: more complex than Montreal, more urgent than Paris, requiring a hybrid architecture.
The protocol's political history is also instructive. Its passage required sustained scientific advocacy, civic mobilization, and political leadership over more than a decade. Industry initially opposed the restrictions, funded uncertainty-promoting research, and deployed the same delay tactics that fossil fuel companies deployed on climate and that technology companies now deploy on AI. The opposition was overcome, not by the objective merits of the case, but by the sustained work of making the case politically unavoidable. Gore's framework insists that AI governance will require similar sustained work, over similar timescales — except that AI is moving faster than ozone depletion moved, which means the timescale for successful intervention is correspondingly compressed.
Origin
The Montreal Protocol emerged from the 1985 Vienna Convention for the Protection of the Ozone Layer, which had established the framework for international cooperation on ozone-depleting substances without imposing binding commitments. The protocol added the binding commitments, triggered by clearer scientific evidence from Antarctic ozone measurements and by advocacy from scientists including Mario Molina, F. Sherwood Rowland, and Paul Crutzen. The United States, under the Reagan Administration, supported the protocol despite its regulatory implications — a political configuration that would be difficult to reproduce in the current environment.
Key Ideas
Successful precedent. The protocol is the empirical demonstration that international governance of transformative technology risks is achievable when specific conditions are met.
Empirical triggers. Connecting scientific measurements to automatic regulatory response avoids the need for continuous political renegotiation.
Alternative availability. Effective regulation requires that compliant alternatives be available when the regulation takes effect; the alternatives enable the political feasibility of strong commitments.
Enforcement mechanisms. Real consequences for defection — trade sanctions, financial penalties — convert normative frameworks into operational governance.
Template for AI governance. The protocol's operational features — triggers, alternatives, enforcement, verification — constitute a template that AI governance debate has been slow to apply despite its availability.
Appears in the Orange Pill Cycle
Governance Requires Condition-Building Work — Arbitrator ^ Opus
The Montreal Protocol's applicability depends entirely on which question you're asking. As a demonstration that international governance of existential technology risks is theoretically possible, the protocol is dispositive (100% — Gore's claim stands). As a template for the operational features that working governance requires — empirical triggers, enforcement mechanisms, verification infrastructure — it is substantially useful (70% — these features translate across domains even when circumstances differ). But as a political roadmap assuming comparable conditions, it is misleading (20% — the contrarian view dominates here).
The critical insight is that the protocol did not succeed because the conditions were favorable. It succeeded because sustained work created the conditions: scientific infrastructure that made ozone depletion measurable, civic mobilization that made inaction politically untenable, industrial repositioning that made alternatives available, diplomatic architecture that converted normative commitments into binding obligations. The protocol's lesson is not "look how easy it was" but "look how much work was required even under relatively favorable circumstances."
For AI, this suggests a re-framing: The Montreal Protocol is not evidence that AI governance will happen, but a specification of the work required to make it possible. Building measurement infrastructure (Climate TRACE for AI deployment patterns). Creating civic demand that political leaders cannot ignore. Developing and demonstrating viable alternatives to uncontrolled deployment. Constructing enforcement mechanisms with real teeth. The protocol's true precedent is not its success but its decade of preparatory labor — which AI governance has barely begun, despite Gore's claim that the timeline is now compressed.
Further reading
- UNEP, Montreal Protocol on Substances that Deplete the Ozone Layer, 1987
- Richard Elliot Benedick, Ozone Diplomacy (Harvard, 1998)
- Stephen O. Andersen and K. Madhava Sarma, Protecting the Ozone Layer: The United Nations History (Earthscan, 2002)
- Susan Solomon, Solving the Climate Crisis (MIT Press, 2024)