Paper 2025/726

Public-Key Quantum Fire and Key-Fire From Classical Oracles

Abstract

Quantum fire was recently formalized by Bostanci, Nehoran and Zhandry (STOC’25). This notion considers a distribution of quantum states that can be efficiently cloned, but cannot be converted into a classical string. Previously, work of Nehoran and Zhandry (ITCS’24) showed how to construct quantum fire relative to an inefficient unitary oracle. Later, the work of Bostanci, Nehoran, Zhandry gave a candidate construction based on group action assumptions, and proved the correctness of their scheme; however, even in the classical oracle model they only conjectured the security, and no security proof was given. In this work, we give the first construction of public-key quantum fire relative to a classical oracle, and prove its security unconditionally. This gives the first classical oracle seperation between the two fundamental principles of quantum mechanics that are equivalent in the information-theoretic setting: no-cloning and no-telegraphing. Going further, we introduce a stronger notion called quantum key-fire where the clonable fire states can be used to run a functionality (such as a signing or decryption key), and prove a secure construction relative to a classical oracle. As an application of this notion, we get the first public-key encryption scheme whose secret key is clonable but satisfies unbounded leakage-resilience (Cakan, Goyal, Liu-Zhang, Ribeiro [TCC’24]), relative to a classical oracle. Unbounded leakage-resilience is closely related to, and can be seen as a generalization of the notion of no-telegraphing. For all of our constructions, the oracles can be made efficient (i.e. polynomial time), assuming the existence of post-quantum one-way functions