Dr. Premchand Bhagwan Ambhore is an accomplished academician and researcher with over 28 years of teaching experience, currently serving as Assistant Professor and Incharge of the Data Centre & Center of Excellence in IIoT at Government College of Engineering, Amravati, India. He holds a Ph.D. in Computer Science and Engineering from Government College of Engineering, Amravati (2015), an M.E. in Computer Science and Engineering from Amravati University (2004), and a B.E. in Computer Science and Engineering from Government College of Engineering, Amravati (1995). His research expertise spans intranet security, cryptography, network security, intrusion detection systems, blockchain technology, cloud computing, and content-based image retrieval. He has guided four Ph.D. students, three M.Tech students, and over 40 undergraduate and diploma projects. He has published extensively in international journals and conferences, authored two books, and holds international and national design patents for intrusion prevention and fiber optic intrusion detection devices. He serves on the editorial boards of several international journals and as a technical program committee member and reviewer for numerous IEEE and Scopus-indexed conferences. He is a member of professional bodies including ACM, IEEE, the Institute of Engineers (India), and the Computer Society of India

Abstract Title: Cryptography for Artificial Intelligence: A Comprehensive Survey Abstract:Artificial Intelligence (AI) has rapidly evolved into a foundational technology across diverse domains, including healthcare, finance, autonomous systems, cybersecurity, and scientific computing. While these advancements have significantly improved automation and decision-making capabilities, they have also introduced unprecedented security, privacy, and trust challenges arising from the extensive use of sensitive data, large scale model training, distributed learning, and cloud-based inference. Traditional security mechanisms are often insufficient to address these challenges, motivating the integration of advanced cryptographic techniques into AI systems. This survey presents a comprehensive review of the rapidly emerging field of cryptography for artificial intelligence. We first examine the security and privacy requirements of modern AI systems and establish a unified threat model encompassing the entire AI lifecycle, from data collection and collaborative training to model deployment and inference. We then propose a taxonomy that categorizes cryptographic techniques according to their roles in securing different phases of AI systems. The survey systematically reviews major cryptographic approaches, including homomorphic encryption, secure multi-party computation, federated learning with secure aggregation, differential privacy, zero-knowledge proofs, trusted execution environments, blockchain-based trust mechanisms, and post-quantum cryptography. For each technique, we discuss its underlying principles, recent advances, practical applications, computational trade-offs, implementation challenges, and suitability for various AI workloads. Special emphasis is placed on the emerging security challenges associated with foundation models and large language models, including confidential inference, prompt privacy, model extraction, secure fine-tuning, watermarking, and verifiable AI services. Furthermore, this survey provides a comparative analysis of existing cryptographic techniques with respect to security guarantees, computational efficiency, scalability, communication overhead, deployment complexity, and applicability to modern AI architectures. We also identify key open research challenges, including efficient privacy-preserving deep learning, scalable encrypted inference, hybrid cryptographic architectures, secure multi-agent AI systems, and quantum-resistant AI infrastructures. By consolidating recent advances across cryptography and artificial intelligence, this survey offers a unified reference for researchers and practitioners while highlighting promising future research directions toward developing secure, privacy-preserving, trustworthy, and resilient AI systems. Keywords: Artificial Intelligence, Cryptography, Privacy Preserving Machine Learning, Homomorphic Encryption, Secure Multi-Party Computation, Federated Learning, Differential Privacy, Zero-Knowledge Proofs, Trusted Execution Environments, Post-Quantum Cryptography, Large Language Models, AI Security