Quantum-secure AI models for secure collaborative development in the context of geopolitical instability

The paper considers the urgent problem of ensuring the security of intellectual property in the distributed development of artificial intelligence in the context of escalating geopolitical tensions. The concept of an AI system architecture integrating quantum cryptography methods, in particular quantum key distribution (QKD), is proposed to protect critical workflow stages of collaborative model development and operation. The innovative approach is to create an end-to-end secure circuit, where AI components not only process data, but also ensure their physical security at the encryption key transfer level, significantly increasing the resistance to compromise of classical infrastructure elements. The study was carried out taking into account the specifics of the development of quantum and AI technologies in the Russian Federation and the existing challenges in the field of information security. The basic principles of building such a system and potential directions of standardization are considered.