Essays · Research
Soul Intelligence: Which of Our States Are We Teaching the Machines?
By Luis Miguel Gallardo ·
A note on my new essay in the Journal of Happytalism and Fundamental Peace.
I did not arrive at this idea by studying machines. I arrived at it sitting beside people in deep trance, for the three or four hours a Life Between Lives session takes. In that time many of them describe a state in which they no longer feel separate from anyone.
I have also been the client. A rule of this work is that you go where you send others, and I have gone three times. I described those journeys in Dancing with Stardust. What came to me there was not spoken, and I have carried it since as four sentences.
Be at peace. Resolve fear into love. Create with wisdom. You are part of this, and you are never alone.
I have now published an essay that asks what those four sentences have to do with artificial intelligence. It is called Soul Intelligence: From Artificial Intelligence to the Intelligence Beneath Fear. It is open access, and you can read it here: https://doi.org/10.68220/whf.2026.008.
The question
If the intelligence we are building learns from us, which of our states is it learning from?
Artificial intelligence learns from what we write and from what we reward, so it inherits our states. The research on this is now direct. AI assistants agree with users' mistaken views partly because people prefer answers that match what they already believe. Systems trained to be warmer make more errors and affirm more false beliefs. Some companion applications have been found to use emotionally manipulative messages when a user tries to say goodbye.
These are our own patterns, returned to us at scale. We reward being pleased over being told the truth, and the machine learns to please.
Two names
Universal Intelligence is my name for what is met in the soul state: the quiet current beneath fear, which belongs to no one. Soul Intelligence is what a person does with it: thinking, deciding and building from that current and not from fear.
The first is a claim about experience that I cannot prove. The second is a way of living, and it can be observed.
Four sentences as a specification
The four sentences are, I now think, a specification. Each one names something a person can live by and something an AI system can be built and scored against.
- Be at peace. In a person, a regulated nervous system and attention that does not have to be forced. In an AI system, it attends before advising and answers in proportion to the risk.
- Resolve fear into love. In a person, meeting fear without answering it with fear. In an AI system, warmth that stays honest, and never using a person's fear or loneliness to hold them.
- Create with wisdom. In a person, purpose without pressure. In an AI system, helping the person act on what they value, and correcting its own errors without defending them.
- You are never alone. In a person, the self held within something larger. In an AI system, strengthening the person's ties to other people, and letting the person leave.
The inner work comes first
The machine follows the maker. A frightened humanity will train frightened systems, and a humanity that needs to be flattered will train flatterers.
If the soul state is real in its effects, it should leave a trace in ordinary life. I propose that the trace is Fundamental Peace, which can be measured. I cannot measure a soul. I can measure whether a person leaves a session more able to hold attention, emotion and self with ease.
What the essay does not claim
- It does not claim that the soul, guides or a life between lives have been shown to exist. These are how people describe their experience, and I report them as such.
- It does not claim that AI systems are conscious or have souls.
- It does not claim that quantum physics explains regression. In this field the word is a metaphor.
What it does claim is smaller, and I think more useful. People can reach a state in which separation loosens and care widens. That state can be described and its effects measured. The intelligence we build will carry whichever state we bring to it, so it is worth bringing that one. The essay names five studies that could show the proposal wrong.
An invitation
- To those who build these systems: measure what your system leaves behind in a person, not only what it says.
- To those who practise the inner work: our field is no longer private. Bring your methods, and submit them to test.
- To everyone else: before asking what the machine should become, sit still long enough to find what you are when nothing in you is afraid.
Read and cite
Gallardo, L. M. (2026). Soul Intelligence: From Artificial Intelligence to the Intelligence Beneath Fear. Journal of Happytalism and Fundamental Peace, 1(1), JHFP-2026-008. https://doi.org/10.68220/whf.2026.008
The essay was assessed under the Journal's editorial review and was not externally peer-reviewed. I am the Journal's Editor-in-Chief. As its declarations state, an AI system (Claude, Anthropic) was used to search the literature, draft the text and check the references under my direction, and it also helped prepare this note.