Threat Modeling of ML-intensive Systems: Research Proposal

Context: The rise of Artificial Intelligence (AI) and Machine Learning (ML) applied in many software-intensive products and services introduces new opportunities but also new security challenges. Motivation: AI and ML will gain even more attention from industry in the future, but threats caused by already discovered attacks specifically targeting ML models are either overseen, ignored, or mishandled. Problem Statement: Current Software Engineering security practices and tools are insufficient to detect and mitigate ML Threats systematically. Contribution: We will develop and evaluate a threat modeling technique for non-security experts assessing ML-intensive systems in close collaboration with industry and academia.