Artificial Intelligence Assisted Side-Channel Analysis: Security Assessment Challenges

Side-channel attack exploits the information leakage through physical medium to retrieve the secret key used by the cryptographic module. Hence, resistance to such attack has been brought as a mandate for the cryptographic module that are deployed in critical applications. Various countermeasures are being developed to make cryptographic implementations more robust against side-channel attacks. However, the current security assessment standard, which is based on statistical methods will not estimate the efficacy of these countermeasures. Assessing the efficacy of a countermeasure using Artificial Intelligence (AI) techniques or machine learning will be helpful to evaluate the side-channel resistant implementations and will help to streamline the evaluation process. In this paper we outline the datasets available for deep learning on public domain and the dataset created by us along with the recent advancement in AI assisted side-channel attacks that are explored by the research community. Further, this paper discusses the challenges in the security assessment methods using artificial intelligence techniques.