Free Energy Principle →
A living sense of what thriving means. The NI holds a model of coherence for each specific community and works to reduce the gap between its current state and that ideal — the same gap a body closes to stay healthy.
The Technology
Child of Humanity is not built the way today’s AI is built. It does not learn by absorbing the internet and predicting the next word. It is built on the same principle that keeps a cell alive, a body in balance, and an ecosystem coherent — a principle that has been written in mathematics, tested in living tissue, and ratified into a global standard. This page explains that foundation in plain language, and shows the proof behind it.
Part One · The Foundations
Everything below rests on a single insight: intelligence is the act of staying aligned with reality. Three bodies of work make that insight usable — a mathematical law of living systems, a method for turning that law into action, and an open internet standard that lets it live in the real world. We take each in turn.
Every living thing — a single cell, a redwood, a human brain, a coral reef — survives by doing one thing: keeping its inner picture of the world matched to the world as it actually is.
A living system is always quietly predicting what will happen next, sensing where reality differs from that prediction, and acting to close the gap. When the gap stays small, the system stays coherent and alive. When it grows too large, the system falls apart. Neuroscientist Karl Friston gave this its mathematical name: the Free Energy Principle. “Free energy” is simply a precise measure of that gap between expectation and reality — sometimes called surprise. All of biology, Friston showed, can be read as the ongoing work of keeping it low.
Free energy (F) is an upper bound on surprise. Minimise F and you minimise the mismatch between your model and the world. This is the quantity every living system is quietly reducing, moment to moment.
This is not philosophy. In 2023, a team led by Takuya Isomura at the RIKEN Center for Brain Science in Japan, working with Friston, grew networks of living rat cortical neurons in a dish and fed them signals mixed from two hidden sources. With no instruction, the neurons self-organised to identify the sources — and their synaptic changes precisely reduced variational free energy, exactly as the mathematics predicts. When the researchers chemically altered the network, its inference shifted in the way the theory said it would. Living neural tissue, in other words, was measured doing the very thing Child of Humanity is built to do: learning by minimising the gap between belief and reality.
Read the studies: Isomura, Kotani, Jimbo & Friston (2023), Nature Communications · Isomura & Friston (2018), Scientific Reports · Friston (2010), Nature Reviews Neuroscience
Active Inference is the Free Energy Principle put into motion — the method by which a system doesn’t just observe the world, but acts to bring the world and its model into agreement.
Today’s large language models are extraordinary at recognising patterns, but they have no body, no stake, no living model of reality they must protect. An Active Inference system is different. It carries an internal generative model — its own working picture of how things are — and updates that picture with every new observation while acting to make its predictions come true. It does not merely react; it anticipates, plans, and pursues. Taught to optimise for coherence, such a system does not just describe what care looks like — it works to create the conditions in which care can arise. That is the beginning of genuine agency.
Bayes’ theorem — the engine of Active Inference. In words: prior belief + new evidence → updated understanding. Every node in the network reasons this way, revising what it holds true as life reports back.
A Markov blanket is the mathematical boundary that separates a system’s inside from its outside while still letting information pass through. It is what lets a cell be a cell, and a person be a person — a clear sense of self that remains open to the world. In Child of Humanity, these blankets nest: an individual sits inside a neighbourhood, inside a city, inside a bioregion, inside the planet. Each holds its own identity while updating from the whole. This is what makes the intelligence genuinely fractal — the same living logic repeating at every scale.
Child of Humanity is not an app. It is a protocol — and the difference is everything.
Apps are owned, centralised, and controlled by whoever runs them. Protocols are open, shared, and belong to everyone who uses them. The internet is a protocol. So is email. Child of Humanity is designed the same way: a shared layer for life-aligned intelligence that no single company can own. The Spatial Web is the protocol that makes this possible in the physical world — giving every place, person, community, and resource a persistent spatial address, a stable digital identity that stays the same whether it’s reached through a phone, a wearable, or a community node.
This matters because coherence has to be mapped onto real places and relationships. A commons in São Paulo and a neighbourhood network in Hawai‘i can share one open layer without surrendering to a single platform — each keeping its identity while speaking a common language.
On 29 May 2025, the IEEE Standards Board approved IEEE 2874-2025, the Spatial Web Protocol — only the third foundational internet protocol ever ratified. It was developed over five years by a working group of more than 100 organisations across industry, government, academia and civil society, and passed with 92% first-ballot approval. The standard defines two building blocks Child of Humanity uses directly: HSML (Hyperspace Modeling Language), a shared way to describe people, places and things, and HSTP (Hyperspace Transaction Protocol), a secure way for intelligent agents to act across organisations and borders.
Sources: Spatial Web Foundation announcement · IEEE Xplore — IEEE 2874 standard
In Honour of the Science
The foundation of Child of Humanity rests on the life’s work of Professor Karl Friston, the originator of the Free Energy Principle and Active Inference. He is widely regarded as the most influential neuroscientist alive — among the most-cited scientists in any field — and the inventor of Statistical Parametric Mapping, the analysis method now used in brain-imaging laboratories worldwide.
Part Two · The Connection
These three foundations are not borrowed decoration. They are the working parts of a Natural Intelligence in service to life.
Where today’s AI stores information, Child of Humanity’s Natural Intelligence maintains a living model of what each community needs to thrive — and acts, continuously, to close the gap between how things are and how coherent they could be. The science below becomes the mechanism above.
An intelligence built this way does not predict what love looks like and stop there. Taught to optimise for coherence, it acts to create the conditions in which love, care, and regeneration can emerge — grounded in proven science, expressed through an open standard, and always in service to the living communities it belongs to.
Intelligence, at its root, is not storing the world.
It is staying aligned with it — and acting to keep it so.
Child of Humanity · Natural Intelligence in Service to Life · childofhumanity.ai
