Credit Scoring Is a Complex Systems Problem
As lenders increasingly automate credit decisions with AI, many of the industry's most important challenges—including fairness, validity, and governance—cannot be solved through machine learning alone.
Why Open Credit Scoring?
Financial services is entering a new era of AI-driven decision making. From machine learning models to generative and agentic AI systems, lenders are rapidly adopting increasingly sophisticated technologies to automate underwriting, improve operations, and expand access to credit.
Open Credit Scoring seeks to establish the scientific foundations for trustworthy AI-enabled credit decisions. Rather than treating fairness, validity, transparency, and governance as isolated challenges, we approach them as properties of a larger complex system that includes people, institutions, policies, and feedback loops.
To understand and improve these systems, we use systems thinking to integrate machine learning, causal inference, and system dynamics into a unified framework for studying AI-enabled credit decisions. Together, these approaches help explain not only how models make predictions, but how AI systems shape—and are shaped by—the social, economic, and legal systems in which they operate.
Research
Proxy Discrimination in Alternative Data
When does alternative data improve credit access, and when does it act as a proxy for protected characteristics?
Develop causal methods to distinguish legitimate predictors of creditworthiness from discriminatory proxies.
Sources of Disparate Impact
What causes disparate impact in AI credit decisions, and how can those causes be addressed?
Develop causal and system dynamics models to identify the sources of disparate impact and evaluate potential interventions.
Trustworthy AI Credit Underwriting
How can lenders responsibly automate credit underwriting using generative and agentic AI?
Develop methods that combine causal, generative, and agentic AI while preserving accuracy, explainability, and trust.
Standards
IEEE P3591: Standard for Fair Decision Making Through Causal Analysis
How can organizations determine whether an AI credit decision is fair, valid, and legally compliant?
IEEE P3591 establishes a common framework for evaluating AI systems by connecting concepts from anti-discrimination law, causal inference, and machine learning.
Mission
Advancing Science and Trust Through Open Collaboration
Open Credit Scoring brings together industry, nonprofits, academia, and government to advance the research, standards, and governance needed for trustworthy AI credit decisions.
Our mission is twofold: to establish the scientific foundations of trustworthy credit scoring and to build the collaborative infrastructure needed to support innovation, competition, and trust across the credit ecosystem.
Industry
Translate Research into Real-World Innovation
Lenders, fintechs, data providers, and technology companies contribute practical expertise, real-world challenges, and implementation experience that help translate research into practical credit solutions.
Nonprofits
Represent Consumer and Community Interests
Civil-rights organizations, consumer advocates, and community groups help identify risks, evaluate impacts on affected communities, and ensure that new approaches promote fair and trustworthy access to credit.
Academia
Advance the Science of Credit Scoring
Researchers contribute new methods in systems thinking, causal inference, machine learning, economics, and law that help establish the scientific foundations of trustworthy credit decisions.
Government
Promote Safety, Soundness, and Consumer Protection
Regulators engage with emerging research, technical standards, and industry practices to help promote safety, soundness, consumer protection, and confidence in AI-enabled decision systems.
About
Chris Lam is the founder and CEO of Epistamai, an AI startup doing research on algorithmic bias in credit underwriting. He also chairs the IEEE P3591 standard. He previously worked as a data scientist at the Federal Reserve Bank of Chicago, a technology strategist at Hewlett-Packard, and a project leader at Consumer Reports. He holds a B.S. in computer science from the University of Pennsylvania, an M.S. in electrical engineering from Columbia University, and an M.B.A. from Northwestern University (Kellogg).
Get Involved
Collaborate With Us
We welcome inquiries from researchers, lenders, regulators, nonprofits, and technology providers interested in advancing the science of trustworthy credit decisions.
Whether you want to collaborate on research, contribute to a standard, or participate in governance discussions, we'd like to hear from you.
Newsletter Sign Up
Receive updates when we publish new research, release draft standards, announce workshops, or open standards for public comment. No marketing.