In today's society, which is heavily influenced by technology, it is crucial to prioritize the security and integrity of computer systems and their underlying hardware components. As advancements in hardware technologies continue to progress rapidly, new vulnerabilities emerge, posing significant risks to the confidentiality and integrity of sensitive information. Therefore, it is essential to proactively identify and mitigate potential threats to hardware. Traditionally, threat modeling tools have primarily focused on software vulnerabilities, neglecting the exploration of hardware vulnerabilities. However, with the increasing complexity of hardware architectures, there is an urgent need for effective methodologies to assess and address potential threats at the hardware level. Currently, state-of-the-art approaches in hardware threat modeling rely on static analysis techniques and knowledge of known hardware vulnerabilities. The existing approaches are considered cumbersome since they require computer validation experts and engineers to perform manual inspection, simulationbased testing, and formal verification. These approaches face increasingly difficult challenges these days when hardware architectures continuously evolve, rendering them more advanced and complicated. In order to cope with these challenges, it is therefore essential to seek alternative and more reliable approaches that are capable of improving the efficacy and accuracy of threat identification and analysis. Since generative Artificial Intelligence (AI) is equipped with the tool to model and generate complex data patterns, it can be considered an alternative approach for hardware threat modeling. With the aid of generative AI, the restricted scope faced by threat modeling can, therefore, be expanded. By incorporating generative AI into hardware threat modeling, hardware vulnerabilities that are hard to be detected and analyzed by conventional approaches can be identified. Hence, the overall security and integrity of computer systems can also be significantly enhanced, resulting in the formation of a more secure environment for protecting sensitive information.
Keywords: Artificial intelligence, Hardware security, RAG approach, RISC-V architecture, Threat modeling.
