INDEPENDENT AI RESEARCH ORGANIZATIONUNITED STATES / SIGNAL EI–001
OBSERVE. TEST. UNDERSTAND.
Independent research for safer intelligence.
We study how AI systems learn and behave. Through controlled experiments, we investigate how their capabilities develop and where their behavior becomes unreliable.
Empirical Institute is an independent research organization dedicated to the empirical study of artificial intelligence.
We use controlled environments and behavioral evaluations to study how AI systems change as they learn. Our questions concern uncertainty, cooperation, adaptation, and failure: what changes, under which conditions, and how can we measure it?
The aim is practical: methods and evidence that help researchers identify safety problems during development.
STATUS ACTIVEMETHOD EMPIRICALPROGRAM AI SAFETY RESEARCHORIENTATION INDEPENDENT
CH 02 / RESEARCH PHILOSOPHY
Study intelligence while it develops.
Much of AI safety evaluates systems after significant capabilities already exist. We examine intelligence as a developing system — observing learning trajectories, uncertainty, cooperation, adaptation, and failure before behaviors become deeply embedded.
01 /
Development
Study how capabilities and behaviors emerge over time.
02 /
Environment
Use controlled simulated environments to observe learning and adaptation.
03 /
Measurement
Identify behavioral patterns, uncertainty, and emerging capabilities.
FOUR CONNECTED RESEARCH PROGRAMS
Four programs, one research agenda.
Our work connects behavioral evaluation, developmental learning, simulated environments, and curriculum design.
What should we measure before intelligence becomes more capable?
“
When does a model begin to recognize uncertainty?
01 / 07
EXPERIMENTAL INTERFACE
Signal study.
An illustration of the observation interface. Illustrative simulation — not live research data.
EI-SIM-001 / CONTROLLED ENVIRONMENTILLUSTRATIVE SIGNAL
AGENTS04
ITERATION01842
SIGNALNOMINAL
UNCERTAINTYOBSERVED
OUR METHOD
Evidence before assumption.
Empirical Institute favors measurable evidence over speculation.
01 —
Observe
Identify behavior or capability worth understanding.
02 —
Isolate
Create controlled experimental conditions.
03 —
Test
Expose intelligent systems to structured environments and tasks.
04 —
Measure
Collect behavioral and performance evidence.
05 —
Understand
Develop hypotheses about why the observed behavior emerged.
CH 07 / PUBLICATIONS
An open record of inquiry.
Research papers, experiments, tools, and observations from our work.
▤
Research archive in development.
Future research outputs appear here as they become available.
ARCHIVE / STANDBY
CH 08 / PEOPLE
Across disciplines. Around a shared question.
Empirical Institute brings together researchers, technologists, scientists, and interdisciplinary thinkers interested in understanding how intelligence develops.
Eric Choi founded Empirical Institute 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.
“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.”
FOUNDING TEAM
Eric Cosentino
Founding team member, Empirical Institute
Mathematical Strategist | Game Theory in AI & Cybersecurity | Adversarial Systems Modeling
Former mathematics teacher bringing a mathematical and strategic perspective to the institute’s research into artificial-agent behavior. His work focuses on applying game theory, adversarial systems modeling, and mathematical reasoning to AI and cybersecurity, including how intelligent agents behave under competition, cooperation, uncertainty, conflicting incentives, and adversarial pressure.
At Empirical Institute, this perspective supports the design of developmental games, multi-agent environments, adversarial experiments, and quantitative frameworks for measuring behavioral change.
SHARED INQUIRY / COLLECTIVE UNDERSTANDING
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.
Empirical Institute welcomes conversations with foundations, research partners, grantmakers, philanthropic organizations, and institutions interested in advancing rigorous AI safety research.