AI Model Security and Robustness
Artificial Intelligence (AI) models are increasingly deployed in critical domains such as cybersecurity, healthcare, finance, autonomous systems, manufacturing, and smart infrastructure, where their reliability and security are essential. However, AI models are vulnerable to a wide range of threats, including adversarial attacks, data poisoning, model evasion, model theft, privacy leakage, and unauthorized manipulation. Ensuring the security, robustness, and trustworthiness of AI models throughout their lifecycle has become a fundamental challenge for researchers and practitioners. Robust AI systems must be capable of maintaining high performance under adversarial conditions while preserving the confidentiality, integrity, and availability of data and model assets.
The International Conference on Cyber Security in Artificial Intelligence (ICCSAI 2027) invites researchers, academicians, cybersecurity professionals, industry experts, and policymakers to present innovative research and practical solutions in AI model security and robustness. This track focuses on secure machine learning algorithms, adversarial machine learning, robust deep learning, AI model verification and validation, explainable and trustworthy AI, privacy-preserving machine learning, federated learning security, secure model deployment, AI supply chain security, model integrity protection, secure MLOps, AI governance, and resilient AI architectures. Contributions addressing the security of foundation models, large language models (LLMs), generative AI, computer vision, natural language processing, reinforcement learning, and autonomous AI systems are particularly encouraged.
This track also explores emerging techniques for detecting and mitigating adversarial attacks, protecting intellectual property through secure model management, enhancing AI resilience against evolving cyber threats, and ensuring compliance with ethical and regulatory frameworks. Participants will discuss innovative methodologies, benchmark datasets, evaluation frameworks, and real-world applications that improve the robustness and dependability of AI systems. Through keynote presentations, technical sessions, and interdisciplinary collaboration, ICCSAI 2027 provides a premier platform for advancing secure, resilient, and trustworthy AI technologies that strengthen cybersecurity and enable the safe deployment of intelligent systems across diverse digital environments.
