Machine Learning and Fraud: A Practitioner Perspective
Machine learning is rapidly reshaping how the payments industry approaches fraud—moving from static, rules-based systems to dynamic, predictive intelligence. But what does that shift look like in practice?
In this episode of the U.S. Payments Forum Podcast, industry experts explore how machine learning is being applied today to detect fraud, reduce false positives, and adapt to evolving threats in real time. From understanding the difference between predictive AI and generative AI to integrating machine learning into existing fraud stacks, this conversation offers a practical, real-world view into what’s working—and what’s next.
Listeners will also gain insight into implementation challenges, governance considerations, and the growing importance of cross-channel data, as well as a forward-looking perspective on emerging threats like scams and deepfakes.
In This Episode, You’ll Learn:
- How machine learning differs from traditional rules-based fraud systems
- The role of predictive AI vs. generative AI in fraud prevention
- How organizations are using machine learning to improve detection and reduce false positives
- Key implementation challenges, including data integration and model governance
- Ethical considerations and bias in machine learning models
- What’s next: emerging threats, deepfakes, and the future of fraud prevention
Host: Marie Jordan, Visa
Guests: Lenny Gusel, Feedzai and Jonathon Robinson
Please note: The information and materials available on this web page (“Information”) is provided solely for convenience and does not constitute legal or technical advice. All representations or warranties, express or implied, are expressly disclaimed, including without limitation, implied warranties of merchantability or fitness for a particular purpose and all warranties regarding accuracy, completeness, adequacy, results, title and non-infringement. All Information is limited to the scenarios, stakeholders and other matters specified, and should be considered in light of applicable laws, regulations, industry rules and requirements, facts, circumstances and other relevant factors. None of the Information should be interpreted or construed to require or promote the establishment of any solution, practice, configuration, rule, requirement or specification inconsistent with applicable legal requirements, any of which requirements may change over time. The U.S. Payments Forum assumes no responsibility to support, maintain or update the Information, regardless of any such change. Use of or reliance on the Information is at the user’s sole risk, and users are strongly encouraged to consult with their respective payment networks, acquirers, processors, vendors and appropriately qualified technical and legal experts prior to all implementation decisions.

