Abstract
AI is rapidly moving from experimental prototypes to business-critical application. Machine Learning, LLMs and Agentic AI are set to transform how companies optimize operations, manage personnel, forecast and influence demand, and design the client journey. In financial institutions, both front-office and internal functions like compliance and risk management are going to be impacted. This panel will explore how AI can bridge the gap between “chat-boxes” and real-world business impact, through managing data flows, addressing privacy and security constraints, ensuring auditability, and integrating AI into decision-making processes. The discussion will focus on trustworthiness and interpretability, rigorous stress-testing under real-world conditions, and the role of causal analysis in avoiding unintended consequences. We will touch upon approaches ranging from classical machine learning to Transformers and AI Agents and discuss best practices for AI governance and AI model risk management.
Target Audience
AI researchers interested in practical impact and responsible deployment; industry practitioners building AI products and decision-support systems; FinTech leaders exploring AI applications in risk, compliance, and investments; advanced students and researchers seeking applied research directions
Presenters
- Massimo Morini (moderator)
- Marcos Lopez De Prado
- Alex Lipton
- Michael X Zhang (TBC)
- Lin Jianwu (TBC)
- Sandy Pentland
Outline and Description of the Panel
- Opening from Moderator
- Differences and common issues across industries and sectors
- Capabilities and limits across approaches
- Trustworthiness, stress-testing, model risk management and governance
- Data flow, privacy, interpretability and auditability
- Decision integration and human oversight
- Finance-specific use cases and issues
- Research gaps and open problems