Lecture
Securing Edge AI by Design: Practical Trust Foundations for Modern Embedded Systems
- at -
- Hall C6C6.501
- Language: English
- Type: Lecture
Lecture description
As embedded systems become more connected, intelligent, and software-defined, developers face a difficult challenge: how to add AI, connectivity, updateability, and long product lifecycles without expanding the attack surface beyond what teams can realistically manage. This is especially urgent for industrial, automotive, medical, consumer, and IoT devices, where security decisions made early in the design process can determine whether a product remains trustworthy, supportable, and compliant over time. This presentation will provide a practical framework for building security into embedded systems from the start, with particular attention to AI-enabled edge devices and the urgency around PQC to protect against tomorrow’s threats today. It will examine the foundational capabilities developers need, including hardware-rooted trust, secure boot, device identity, key management, secure debug, authenticated update mechanisms, vulnerability handling, impact of quantum and lifecycle resilience. It will also address how these foundations support emerging requirements around embedded security, functional safety, and regulatory readiness without creating unnecessary cost or development complexity. The session will focus on practical engineering decisions rather than product promotion. Attendees will learn how to evaluate security trade-offs across hardware, firmware, operating systems, connectivity, and AI workloads; how to design systems that can be updated and defended after deployment; and how to avoid treating security as a late-stage compliance exercise. Key takeaways will include a security-by-design checklist for embedded teams, guidance on aligning security architecture with cost and lifecycle constraints, and examples of how platforms, components, and tools can be used more effectively to build trustworthy connected devices.
