Humanity at the center: ethical AI and the water-energy-food nexus (HLPF side event)

Session: “Humanity at the Center: Ethical AI, the Water-Energy-Food Nexus, and Global Partnerships for Transformative SDG Action” (Conference Room 3, 1:15 to 2:30 PM)
Hosts: The Permanent Mission of the Republic of Korea to the UN, with IVECA Center for International Virtual Schooling and the Incheon Metropolitan City Office of Education
Keynote: Kaveh Madani (Director, UN University Institute for Water, Environment and Health, UNU-INWEH)
Format: Opening remarks, keynote, and project presentations by secondary and university students from Korea, China, Cameroon and the United States

AI is a physical product

The keynote from Kaveh Madani of UN University did the useful work of moving AI out of the metaphor of the cloud. The public treats AI as something weightless and digital. It is not.

Kaveh Madani, UNU-INWEH: “AI is not a virtual thing; it’s a physical product in a way.”

He tied it to a concrete UN action. A UNU-INWEH report on AI’s water, carbon and land footprint fed the Secretary-General’s AI Environmental Transparency Initiative, launched in June, which calls on AI companies to disclose those footprints. The point of the nexus framing is that a gain in one resource is rarely free in the others, and Madani refused the instinct to sort technologies into good and bad.

Kaveh Madani: “AI is just like a knife. With a knife, you can save a patient’s life in the operating room, or you can kill someone as a murderer. The way we use a technology determines if that technology is our friend or enemy.”

The efficiency paradox, and the farmers who hold the water

Madani’s warning about efficiency is the one worth carrying out of the room, because it undercuts the standard techno-optimist answer. Making a system more efficient makes it cheaper, which makes people use more of it.

Kaveh Madani: “When efficiency improves, things become more affordable. That means that use increases, so we will have more and more consumers, more users, and that means that all those efficiency gains are neutralized and are canceled out.”

On water, he restated the framing he had used at the Forum’s SDG 6 review three days earlier, that the world has entered an era of water bankruptcy, and located the leverage where it actually sits, which is not in the tech sector at all.

Kaveh Madani: “70% of the world’s water is in the hands of farmers, and they are the ones who control water.”

Students found the same problem independently

The most telling detail was structural. Working through IVECA’s cross-border virtual classrooms, student teams in Korea and China arrived separately at the same subject: the water and energy load of the AI data centres that everyone else in the building was celebrating. Their proposals ran to waste-heat reuse for smart farms, undersea and immersion cooling, and, notably for this site, a real-time resource-metering scheme that would tax data centres on measured water and power rather than estimates. The framing that held the day together was offered as a design principle rather than a slogan.

Dr. Brian Mujas, Global NGO Executive Committee: “Technology has no values of its own. It inherits ours.”

Why it matters for the SDGs

The event connects SDG 6 (water), SDG 7 (energy), SDG 2 (food) and SDG 9 (the technology itself), with SDG 4 running underneath through the education model that produced the student work. The measurement argument is the one to keep. AI’s environmental cost is real, large and largely unreported, which is why the transparency initiative Madani cited is a counting exercise before it is a climate one. You cannot govern a footprint you refuse to measure, and at present most of the industry does not measure it at all.

Watch & read

UN Web TV, recording of the side event (10 July 2026).
UN University Institute for Water, Environment and Health, source of the AI footprint report.
Together for AI for Good, the Geneva keynote on the same tension between AI’s promise and its cost.
Digitalization, statistics and AI · Full HLPF 2026 coverage.

Quotations are lightly edited from an automated (Otter.ai) transcript of the UN Web TV recording and should be read as close paraphrase. Speaker names were reconciled against public records; the transcript garbled most of the hosts and student presenters, who are therefore cited by role or institution. Figures cited in the student presentations were vendor or project claims and were not independently verified.