2026
Multimedia, RAG, and LLMs in High-Security Environments - a reference implementation for the AI2MMIR architecture
The paper concludes that the convergence of trustworthy AI, robust multimedia analysis, and privacy-preserving architectures is essential for responsible AI adoption in regulated domains, and highlights open challenges in cross-document synthesis, multimodal robustness, and scalability for future research.
Stefan Wagenpfeil, Angelina Chernikov
· TRI-IVIS@AVI · 1 citation