Principle 01
Intelligence must respect limits.
The atmosphere, energy systems, material supply chains, public budgets, and human attention are finite. AI should be designed with those constraints visible from the beginning.
A public commitment
The Sustainable Intelligence Manifesto
AI can expand human knowledge without treating planetary systems and communities as disposable inputs.
Principle 02
Efficiency is a capability.
Useful work per unit of energy, hardware, data, and time is a core measure of technical progress—not a secondary optimization.
Principle 03
Knowledge should compound.
Open methods, reusable datasets, shared benchmarks, and reproducible evidence prevent wasteful repetition and expand who can contribute.
Principle 04
Accountability must be measurable.
Environmental and social claims require transparent methods, independent review, and reporting across the complete system lifecycle.
Principle 05
Participation shapes legitimacy.
Communities affected by AI systems must have meaningful authority in deciding whether and how those systems are developed and deployed.
Principle 06
The future is a design stakeholder.
Technical decisions made today must preserve choices, resources, and healthy institutions for future generations.
From declaration to practice.
Read the research, test the standards, and hold our work to the same principles.