Security Risks of AI Hardware for Personal and Edge Computing Devices

This technical report presents an analysis of the potential security benefits and risks associated with hardware for Artificial Intelligence (AI) on personal and edge computing devices. Conducted by Google in collaboration with NCC Group, the report assesses the global landscape from both product development and regulatory perspectives. It employs a threat modeling approach to evaluate the risks to stakeholders, including AI companies, developers, and users, based on different modes of AI integration. Key questions addressed include the security implications of AI on edge devices, the risks of security mono-culture, and the potential for increased exposure to malware attacks. The report also examines the differences between open and closed approaches to AI integration, highlighting the complexities of the attack surface in both edge and cloud computing environments. Furthermore, it discusses compliance with regulations such as the EU AI act and offers strategic recommendations for deploying AI capabilities securely. The findings aim to inform stakeholders about the cybersecurity landscape as AI technologies evolve.