Intelligence is something
to observe.
Empirical Institute is an independent research organization studying the development, evaluation, and safety of intelligent systems.
What We Study
We investigate how artificial intelligence learns, adapts, reasons, behaves, and develops capabilities over time, with the goal of building empirical methods for understanding and improving AI safety. Developmental AI safety asks how behaviors emerge during learning, and which experimental conditions help us characterize those changes.
Why Empirical?
Questions about artificial intelligence are often philosophical, theoretical, or speculative. Empirical Institute is built around a simpler principle: when possible, test the question.
We seek measurable evidence about how intelligent systems learn and behave. Experiments cannot answer every question about AI, but they can constrain speculation and reveal patterns that theory alone may miss.
Our Philosophy
Intelligence is not merely something to build. It is something to observe.
We favor empirical experimentation, falsifiable hypotheses, reproducibility, and open inquiry. Controlled environments allow us to isolate variables; developmental learning lets us study change over time. Uncertainty is a feature of scientific inquiry to characterize carefully, rather than a reason to overstate our conclusions.
Continuity of inquiry
Empirical Institute carries forward the research philosophy and program structure developed under the Agnostic Ally concept: a commitment to observation, experimentation, and the careful study of developing intelligence.
Evidence before assumption.
Empirical Institute favors measurable evidence over speculation.
Observe
Identify behavior or capability worth understanding.
Isolate
Create controlled experimental conditions.
Test
Expose intelligent systems to structured environments and tasks.
Measure
Collect behavioral and performance evidence.
Understand
Develop hypotheses about why the observed behavior emerged.
Research is a
collaborative process.
We are interested in collaborating with researchers, universities, laboratories, foundations, technologists, and institutions working on artificial intelligence, machine learning, AI safety, developmental learning, evaluation, simulation, and related fields.
Independent inquiry takes long-term support.
Empirical Institute welcomes conversations with foundations, research partners, grantmakers, philanthropic organizations, and institutions interested in advancing rigorous AI safety research.