Unpacking the Threat: AI Model Deserialization Attacks and Defense Strategies

This talk explores deserialization attacks targeting AI model files. We will discuss the mechanics of deserialization, review common AI model file formats, and identify those most vulnerable to exploitation. Practical guidance on mitigating risks by choosing secure formats and implementing robust practices will be provided. By shedding light on this emerging threat, this talk aims to empower developers and security professionals to build more resilient AI systems.

Speaker: Austin Coleman is a Cyber Security Engineer with six years of experience in the field. He has his undergraduate degree in Computer Science from the University of California, San Diego, and has worked across various domains in security, including architecture, operations, incident response, and threat intelligence. Recently, he's focused on the emerging field of AI security.