The climate crisis keeps accelerating. The last 11 years were the hottest ever recorded. Scientists now expect a temporary overshoot of the 1.5-degree limit. We must keep it as short and small as possible. By drastically reducing emissions this decade. And by accelerating the transition away from fossil fuels to renewables.

Every fraction of a degree matters. And so does every hour of warning. As we meet, an El Niño is gathering strength in the Pacific. Floods, storms, wildfires and droughts will strike even harder. Yet the world remains woefully unprepared.

Many countries still lack the most basic protection — systems that warn of approaching danger. Hundreds of millions of people are exposed, overwhelmingly in the Least Developed Countries and Small Island Developing States. It is a profound injustice that those who did the least to cause this crisis are the least shielded from its impacts.

Adaptation is a lifeline. Yet adaptation finance remains a fraction of what vulnerable countries need. Developed countries have a duty to close that gap — with support that matches the scale of the crisis.

I launched the Early Warnings for All initiative to help ensure every person on earth is protected. The same objective is promoted by China’s action plan on early warning for climate change adaptation, and by this fantastic AI instrument of the CMA, that is called MAZU.

Early-warning systems are the most cost-effective protection against climate disaster. Where coverage is comprehensive, disaster deaths are at least six times lower.

A timely warning is the difference between an evacuation and a tragedy. 128 countries now have multi-hazard early-warning systems — more than double since 2015. But one-third remain unprotected.

We urgently need to accelerate — and artificial intelligence can help. It can deliver faster, earlier warnings, and alerts people can receive, understand and act on. A city that opens cooling centers before the heatwave peaks. A fishing village that brings its boats ashore. A farmer who plants differently after being warned the rains will fail.

To succeed, technology must be rooted in equity. I see three tests for AI in early warning.

First — observations and data.

No matter how powerful technology becomes, artificial intelligence cannot compensate for missing data. The entire African continent has fewer stations meeting the standards of the Global Basic Observation Network than some individual European countries. Without these basic observations, forecasts suffer and lives are at risk.

Our Systematic Observations Financing Facility exists to close these gaps. And weather and climate data must be treated as a global public good.

Second — capacity and trust.

The countries most vulnerable to climate disasters often have the least access to these technologies. This must change. They need the capacity to develop, operate and sustain these systems.

Warnings must reach every person at risk. In languages they understand. Through channels they can trust.

Third — responsibility.

Artificial intelligence must help solve the climate crisis — not deepen it. Its appetite for electricity, water and land is growing fast.

I have called on every major AI company to power its data centers with renewable energy by 2030, and to disclose its full environmental footprint.

Technology aimed at protecting people must also protect the planet.

Early Warnings for All demands a step change in international cooperation, for the protection of people and property.

China has been an important partner. Its platforms, satellites and AI models are already helping dozens of countries strengthen early-warning systems.

These tools are built with other developing countries, as partners — drawing on local data, so nations can issue their own warnings. And the tools unveiled here today will widen that reach further.

This is the model the world needs — with technology transfer, joint research and local capacities helping countries protect their people.

Early-warning systems are an economic imperative. A development imperative. A security imperative. And above all, a matter of climate justice.

The technology is here. I was amazed today with the demonstration I had at MAZU in the other building, where the Conference was opened.

The partnerships are growing, and we have seen it here in this Conference.

Now let us match them with the urgency and investment this moment demands. And deliver early warnings for all, everywhere.

(Excerpts of United Nations Secretary-General Antonio Guterres’ remarks to the “Dialogue on Early Warnings for All in Response to Climate Change” at the World Artificial Intelligence Conference Meteorological Forum in Shanghai, China on 17 July 2026.)