SPETViD

3rd International Workshop on Security and/or Privacy Enhancing Technologies for Multimodal Data, held in conjunction with the 21st International Conference on Availability, Reliability and Security (ARES). Multimodal data spans text, images, audio, video, sensor streams, and increasingly AR/VR and 3D content, and its growing use in analytics and automation raises substantial security and privacy risks across the full data lifecycle. SPETViD 2026 focuses on practical and trustworthy protection of multimodal data and AI systems, covering Privacy-Enhancing Technologies (PETs), secure-by-design AI practices, provenance and integrity mechanisms, and governance approaches for domains such as smart communities, healthcare, finance, critical infrastructure, and next-generation multimedia systems. The workshop especially highlights emerging issues around multimodal foundation models, retrieval-augmented generation, deepfakes, adversarial attacks, leakage, and distributed or edge-based multimedia processing.