CAIRFI and Capital One announce 2026-2027 AI Research Awards

The Columbia Center for AI and Responsible Financial Innovation (CAIRFI) and Capital One are proud to announce the recipients of the 2026-2027 AI research awards.

Now in its third year, this collaboration continues to build on the mission of advancing responsible AI and innovation in financial services. By combining Columbia University’s world-class academic research with Capital One’s industry-leading expertise, CAIRFI aims to solve complex challenges in financial services through cutting-edge AI.

This year, the center issued a Call for Proposals seeking faculty AI research focused on several key pillars of modern financial technology:

• Continual learning
• Reinforcement learning
• Guardrails and jailbreaking in large language models
• Responsible and trustworthy AI; AI security
• Causal inference and causal machine learning
• Multi-agentic LLMs for complex reasoning
• Reasoning in sequential decision settings
• Multimodality beyond vision and language: LLMs + tabular, LLMs + time series
• Generative AI for synthetic financial data
• Differential privacy, and its applications to information transmission
• Other innovative methods or ideas

AI Research Award Recipients of the 2026 to 2027 Academic Year

Following a rigorous selection process, the following faculty and research projects have been awarded support for the upcoming academic year.

Micah Goldblum

Uncertainty Estimates for Black-Box Application Programming Interface Language Models
As financial institutions increasingly rely on third-party APIs for large language model (LLM) capabilities, understanding when to trust a model’s output is critical. Professor Goldblum’s research focuses on developing robust methods to estimate uncertainty in “black box” models. This work ensures that AI-driven financial insights are backed by measurable reliability, reducing the risk of overreliance on incorrect model outputs.

Adam Block

Understanding and Stabilizing Reinforcement Learning Dynamics in Post-Training Language Models
Reinforcement learning (RL) is a cornerstone of aligning LLMs with human intent, yet the dynamics of post-training can be unstable. Adam Block is investigating the mathematical foundations of these dynamics to create more stable and predictable AI behaviors. His work aims to ensure that models remain safe and effective as they are refined through human feedback and real-world interactions.

Kathleen McKeown

Fast Controllable Update Summarization
In the high-velocity world of financial data, stakeholders need tools that can synthesize new information without losing context. Professor McKeown is developing advanced techniques for “update summarization.” This project focuses on speed and user-specified control, allowing systems to provide accurate, concise updates on shifting market trends or internal data while maintaining high fidelity to the original source.

Lydia Chilton

Agentic LLM Systems for Reliable Regulatory Compliance and Policy Navigation in Financial Services
Regulatory landscapes in finance are dense and ever-changing. Assistant Professor Lydia Chilton is exploring the use of agentic AI—systems capable of autonomous reasoning and action—to help navigate these complexities. This research aims to build reliable AI assistants that can proactively ensure compliance and guide employees through intricate institutional policies.

Baishaki Ray

Secure Code LLM With Reasoning
As AI models are increasingly used to write the code that powers financial infrastructure, security must be built in by design. Professor Ray’s research enhances LLMs for code generation by integrating advanced reasoning capabilities. By grounding code generation in security logic, this project aims to proactively identify and mitigate vulnerabilities during the development of financial software.

Junfeng Yang

Jailbreaking the Jailbreaks: Proactive Defenses for Large Language Models
Adversarial attacks, such as “jailbreaking,” represent a significant security hurdle for the responsible deployment of LLMs. Professor Junfeng Yang is developing proactive defenses that anticipate and neutralize these attacks before they can compromise system integrity. This work is essential for maintaining the safety and trustworthiness of AI systems in high-stakes financial environments.

To learn more about the partnership between Capital One and Columbia University, visit the CAIRFI homepage.