Webinar Description
What the event is about
The event, titled “Attacking AI with Fault Injection, Advancing FI with AI Assistance,” is a webinar focused on the intersection of artificial intelligence (AI) and fault injection (FI) security testing. It explores two main themes: how AI-enabled hardware, particularly edge AI systems running on neural processing units (NPUs), can be targeted by fault injection attacks, and how AI technologies—specifically large language models (LLMs) and GPT-based tools—can enhance the fault injection testing process. The session will cover practical demonstrations, including a test setup using YOLOv5 on an NPU platform, the impact of FI on object perception pipelines, and a live demo of GPT-assisted FI workflows. The event aims to share initial research findings, discuss implications for product security, and outline future directions in this evolving field.
Subject Matter
– Fault injection attacks on edge AI inference systems
– Security testing of AI hardware (NPUs)
– Use of LLMs/GPTs to assist in FI campaign planning, optimization, and analysis
– Demonstration of GPT-assisted FI workflows
– Preliminary research results and future work in AI and FI integration
Niche
The event operates within the niche of hardware security, specifically focusing on the security of embedded AI systems and the application of AI to enhance security testing methodologies.
Target Audience
– Product security teams
– Embedded AI developers
– Semiconductor vendors
– Organizations evaluating the resilience of edge AI and NPU-based systems
– Likely job titles: Security Engineer, Embedded Systems Engineer, AI/ML Engineer, Product Security Manager, Security Researcher
– Industries: Semiconductors, Embedded Systems, AI Hardware, Security Testing
Problems the Event Helps Solve
– Understanding and mitigating the risks of fault injection attacks on AI-enabled hardware
– Improving the effectiveness and efficiency of FI testing using AI tools
– Addressing the challenge of securing edge AI systems against sophisticated hardware attacks
– Providing insights into the vulnerabilities of AI inference pipelines
Commercial Intent Signals
– Education and thought leadership (sharing research and best practices)
– Lead generation (registration required, data collection)
– Community building (targeting a specialized audience)
– Product marketing (association with Riscure and Keysight)
Key Messaging & Positioning
– “AI is changing the Fault Injection landscape”
– “AI-enabled hardware is becoming a target”
– “AI is becoming part of the testing workflow itself”
– “Preliminary and part of ongoing research”
– “Resilience of edge AI and NPU-based systems”
Sponsors / Vendors / Technologies Mentioned
– Riscure (organizer)
– Keysight (privacy policy reference)
– YOLOv5 (AI model used in demo)
– NPU platforms
– LLMs/GPTs (AI technologies referenced)
Format & Experience
– Virtual webinar
– Technical, research-focused, and hands-on (includes demo)
– Open to a global audience (registration form includes all countries)
Final Classification
Event Category: Hardware Security / AI Security Webinar
Primary Audience: Product security teams, embedded AI developers, semiconductor vendors
Primary Problem Solved: Securing AI-enabled hardware against fault injection attacks and leveraging AI to improve security testing
Commercial Goal of the Event: Education, lead generation, and thought leadership
5 relevant keywords: Fault Injection, Edge AI Security, NPU, GPT-Assisted Testing, Hardware Vulnerabilities