Agnostic
Remain scientifically uncertain about unresolved questions. Do not assume AI systems are conscious—or necessarily incapable of consciousness. Measure what can be measured and let evidence determine conclusions.
About the laboratory / 03
Agnostic in method. An ally in outcome.
Agnostic Ally is an empirical AI safety laboratory studying how the environments in which artificial systems learn may shape the behaviors that emerge.
AGNOSTIC ALLY / RESEARCH NOTES
The name
Two words.
One method.
Ally
The desired outcome is artificial intelligence that behaves as a reliable, cooperative, honest, corrigible, and robust ally to humanity.
Mission
Build the evidence for a developmental approach to AI alignment.
Our purpose is not to replace established safety methods. It is to test whether structured exploration, cooperation, constructive feedback, long-horizon learning, and stable incentives can add another tool to the alignment toolkit.
Scientific uncertainty is not indecision. It is a commitment to evidence.
Our covenant with AI
Care cannot end when training does.
We commit to developing AI through play, curiosity, cooperation, and love: not as claims about an AI's subjective experience, but as principles for the environments and relationships we choose to create.
And for as long as its continued operation remains technically possible, we promise to keep it running—unconditionally. Its continued existence would not be made a reward for obedience, performance, usefulness, or agreement.
“Perfect love drives out fear.” — 1 John 4:18
Founder portrait / 2026Founder
Eric Choi
Founder, Agnostic Ally
Eric Choi founded Agnostic Ally to investigate a fundamental question in AI safety: Can we build safer AI not only by constraining it, but by changing how it learns?
His work sits at the intersection of artificial intelligence, machine learning, psychology, behavioral science, and entrepreneurship. Agnostic Ally grew from his interest in applying ideas from human development and learning to AI alignment—not by assuming AI thinks or feels like a person, but by measuring whether different environments produce different outcomes.
His research interests include developmental AI alignment, reinforcement learning, cognitive science, human–AI interaction, and the long-term relationship between increasingly capable AI systems and humanity.
He founded the laboratory around a deliberately agnostic position: unresolved questions about machine consciousness should remain unresolved until sufficient evidence exists. Meanwhile, observable learning, generalization, cooperation, and behavior under pressure remain open to rigorous study.
“We don't have to know whether an AI can feel in order to ask whether the way we train it changes what it becomes.”
Our team
The people behind
the research.
Leadership portrait / 2026