International Association
for Cryptologic Research

We introduce a new variant of malicious obfuscation. Our formalism is incomparable to the existing definitions by Canetti and Varia (TCC 2010), Canetti et al. (EUROCRYPT 2022) and Badrinarayanan et al. (ASIACRYPT 2016). We show that this concept is natural and applicable to obfuscation-as-a-service platforms. We next define a new notion called auditable obfuscation which provides security against malicious obfuscation. Finally, we construct a proof of concept of the developed notions based on well-studied theoretical obfuscation proposals.