explainability-attack

This paper presents the first model extraction attack on graph classification under strict black-box constraints, exploiting subgraph explanations to estimate decision boundaries. The findings reveal that mandated explainability interfaces create exploitable security vulnerabilities in Graph Neural Network services.