UF researchers develop new training method to help AI tools learn safely

A new paper from UF and Visa Research introduces a training method designed to prevent AI models from memorizing sensitive information—a growing privacy risk in modern machine learning systems.