A new open-source project called Zil Lean brings a domain-specific language (DSL) for Lean4, based on Google Zanzibar’s relationship-based access control model. It aims to help AI projects manage complex knowledge graphs with precise, verifiable rules. The tool is available on GitHub under the MIT license, targeting developers who need fine-grained control over data relationships. This approach could make AI systems more auditable by encoding access and inference logic in a formally verifiable language.
Zil Lean is exactly the kind of tool we need right now. AI is racing ahead, but our ability to control and understand these systems is lagging. This DSL gives developers a way to encode knowledge graph rules in Lean4, a language designed for formal verification. That means we can actually prove that our AI's reasoning is correct and secure.
Google Zanzibar already proved that relationship-based access control scales. By bringing that model to Lean4, Zil Lean bridges the gap between high-performance AI and rigorous logic. This isn't just a tool for today. It's a foundation for tomorrow's trustworthy AI. I'm excited to see how the community builds on this.