Cornell Tech has launched a research collaboration with Mastercard to advance methods for evaluating and auditing generative artificial intelligence systems. Supported by Mastercard's AI Governance program, the project will examine how to improve transparency and real‑world benefits from generative AI.
The research effort includes a core team of Cornell Tech faculty members - Allison Koenecke, Angelina Wang, Helen Nissenbaum, Mor Naaman and Nikhil Garg - who are leaders in the fields of algorithmic fairness, AI evaluation, the ethics of technology and human-AI interaction.
"We are excited to join forces with Mastercard on this initiative," said Naaman, associate dean for faculty affairs and the Don and Mibs Follett professor at Cornell Tech, the Jacobs Technion-Cornell Institute, and the Cornell Ann S. Bowers College of Computing and Information Science. "Their genuine dedication to responsible AI, combined with our team's leadership in identifying and addressing the field's core challenges, positions us to make a meaningful impact. We hope to deliver insights that will be directly useful to Mastercard, as well as anyone who cares about responsibly implementing GenAI."
The initiative comes as organizations are assessing how generative AI systems perform in practice, particularly in areas where uneven results may have meaningful consequences, such as financial services, hiring and health care. The collaboration will aim to better understand how to address the needs of diverse users, contexts and deployment environments.
The research will focus on developing more realistic evaluation methods; creating auditing frameworks that detect outcome variance across the AI pipeline; and examining system‑level factors that affect model performance in real‑world use.
As part of the partnership, Mastercard researchers will engage closely with Cornell Tech faculty, sharing industry insights and collaborating on emerging research questions as the work evolves. Cornell Tech will publish related academic findings.
"Mastercard is committed to operationalizing AI governance at speed and scale. By advancing rigorous evaluation and assurance methods, we are strengthening the robustness and reliability of generative AI across real-world use," said Andrew Reiskind, Mastercard's chief data officer. "Our work is informed by ongoing collaboration with leading academic institutions to bring cutting edge research into practical governance frameworks."
Grace Stanley is a staff writer-editor for Cornell Tech.
