Your AI agents might have more access than they need, and that’s a recipe for data leaks. Traditional ACLs can’t keep up with contextual, document-level permissions in RAG or agentic systems. In his NDC Copenhagen talk, Ashish Jha makes the case for Fine-Grained Authorization—showing how tools like OpenFGA and LangChain can enforce strict data boundaries, isolate tenants, prevent leakage, and maintain audit trails, all while handling billions of access checks per second.
Whether you’re cooking up an internal copilot or a customer-facing AI, mastering FGA is your must-win security battle.
Watch on YouTube
Top comments (0)