AI evaluation lacks the measurement science that other fields rely on.
Finance has accounting; quality has metrology. AIMS is an independent research institute building the methods, tools, and open resources to close the gap — and testing that science against real enterprise problems in China.
Better evaluation starts with better measurement.
A model scores 87% on a benchmark. Against what baseline? With what sampling error? Measuring which capability, exactly? Psychometrics, metrology, and educational testing spent the last century answering these questions — AI evaluation can stand on their shoulders.
The AIMS approach: define constructs before collecting data; attach uncertainty to every score; design benchmarks for durability. Research supplies the theory; tools and open resources put it into practice — and real enterprise problems are where the science gets tested.
Four disciplines, one integrated science.
Each pillar answers a core question in measuring AI well — three from science, one reaching into management.
Measurement foundations
Define what is being measured, how to represent it, and where metrics diverge from the construct.
Evaluation science
Protocols that are comparable, decision-relevant, and built to last.
Statistical discipline
Uncertainty, sampling, and distribution shift as first-class citizens of every evaluation.
Governance & management
Wire measurement results into enterprise roles, processes, audit trails, and compliance.
A small core, an open network.
The institute is driven by a small core team, collaborating with researchers and enterprise practitioners as an open network.
Chris
Chris is the founder of the AIMS Institute. He has long worked at the boundary between AI capability evaluation and enterprise adoption, and started AIMS on a simple conviction: rigorous measurement is not just an academic concern — it is the precondition for enterprises to trust AI with their core business.
TAO TAO
TAO TAO is the institute's collaborator, driving the research agenda and open infrastructure alongside Chris. His focus is the validity of evaluation methods in real enterprise settings: when a score moves, does the business outcome move with it?
Many of the institute's questions can only be answered by researchers and practitioners together. If you own AI evaluation, governance, or adoption inside an enterprise, this network is for you.
“When you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind.”
Open by design.
The institute runs with the standards of a research lab and grows like an open-source project. The only entry ticket is a real question.
Follow the work.
Research, tool releases, and events, delivered as they ship. No spam, unsubscribe anytime.
Talk to us about evaluation.
Choosing a model, preparing a launch, or proving impact to leadership — bring us the real scenario.