Skip to main content
NORC at the University Chicago
AI is easy to use but hard to trust. The NORC Center on AI & Data Quality (AIDQ) helps determine when that trust is earned.

We ensure that the data behind AI is sound and that the information produced with AI is defensible enough to inform society’s most important decisions. We draw on methodologists, statisticians, data scientists, and subject-matter experts from across NORC, and apply methods that have been proven by decades of trustworthy research.

Many assessments of AI come from people with a stake in the answer. We take an independent stance toward the tools we assess, including the ones we build ourselves. As researchers, we follow the evidence. We say plainly what holds up and what does not, so the answer you act on is one you can defend.

Our Work

Ensuring AI deployment in regulated settings holds up to scrutiny.

In federal and state government, public health, education, and justice, an AI system is not judged by a demo. It is judged by whether its outputs survive an auditor, a legislature, or a regulator. NORC has spent decades doing the methodological work that lets data and evidence hold up in exactly those settings, and we bring that work to AI.

For the technology companies building these tools, that makes us the partner who knows what “good enough” means inside a regulated program and how to gain adoption for your capabilities.

Recent Work

Artificial Intelligence for Enhancing Data Quality, Standardization & Integration

Applying innovative data science methods to create datasets that support evidence-based decision-making

Client:

National Center for Science and Engineering Statistics

Clinical Decision Support Innovation Collaborative

Advancing patient-centered clinical decision support through innovation, real-world testing, and collaboration

Client:

Agency for Healthcare Research and Quality

California DSS Qualitative Training

Transforming complex, decentralized data into evidence to improve a large state program assisting low-income families

Client:

California Department of Social Services (CDSS)
More Projects

Shaping the Field

The standards for trustworthy AI are a work in progress, and we’re helping establish them.

The norms for how AI is used in research and public service are not yet settled. Methods for evaluating AI outputs, preparing data for AI, and judging whether a system is fit for a given purpose are still taking shape, and the choices made now will shape the field for years.

NORC is positioned to help: we convene across government, research, and technology, develop methods and frameworks others can build on, and keep the work grounded in whether AI serves the public good.

Convenings

We bring government, research, and technology together to work through what trustworthy AI requires.

Frameworks

We develop methods and standards others can adopt, contributing to a field still defining its norms.

Independence

We evaluate AI without selling it, so the standards we help set aren’t tied to any vendor’s product.

Our Experts

AIDQ experts come from survey methodology, statistics, data science, and AI engineering, with applied experience testing AI where it has to be right.