Securing the Future of AI: Authorization Strategies for RAG Systems with LangChain4J and OpenFGA
Conference (INTERMEDIATE level)
As Retrieval-Augmented Generation (RAG) systems become more prevalent in modern AI applications, ensuring robust authorization frameworks is critical to maintain data privacy, security, and compliance. This talk explores the unique challenges and best practices in designing secure authorization mechanisms tailored for RAG systems. We will dive into real-world scenarios where sensitive data is at risk and examine how Role-Based Access Control (RBAC) and Relationship-Based Access Control (ReBAC) using OpenFGA can be implemented in RAG architectures with LangChain4J.
Attendees will learn how to secure the interaction between the retrieval components and generative models while maintaining the system’s efficiency and accuracy. Whether you're an AI architect, security professional, or developer, this session will provide actionable insights to safeguard the future of AI-driven applications.
Attendees will learn how to secure the interaction between the retrieval components and generative models while maintaining the system’s efficiency and accuracy. Whether you're an AI architect, security professional, or developer, this session will provide actionable insights to safeguard the future of AI-driven applications.
Deepu Sasidharan
Okta
Deepu K Sasidharan is a Software Engineer by passion and profession. He is a Java Champion working as a Staff Developer Advocate at Okta. He is the co-chair of JHipster and the creator of KDash and JDL Studio. He is a polyglot programmer working with Java, Rust, JavaScript, Go, and so on. He is also a cloud technology advocate and an open-source software aficionado. He has authored books on Full-stack development and frequently writes about Java, Rust, JavaScript, Go, DevOps, Kubernetes, Linux, and so on, on his blog.