AI Security in System-on-Chip (SOC) Architectures: Fortifying the Future of Embedded Intelligence (M02c)
Securing AI in SOCs is essential for protecting embedded intelligence from evolving threats. Safeguarding model integrity and confidentiality, particularly in edge computing, requires a dedicated strategy to address data breaches, model tampering, and side-channel attacks. This strategy must adhere to evolving regulatory demands for data privacy, cybersecurity, and trustworthy AI while spanning the entire AI life cycle—from model development to integration and usage. This talk explores the challenges and opportunities for securing embedded intelligence in a compliant and future-ready manner.