AdaCore has released GNAT Foundry: Intersection, an open-source demonstrator illustrating how formal methods and trusted verification tools can serve as guardrails for artificial intelligence in high-integrity software engineering workflows.
While artificial intelligence can automate tasks and boost engineering efficiency, it presents distinct challenges in safety-critical development due to unpredictable failure modes and the overwhelming volume of generated output. The GNAT Foundry workflow addresses these issues by applying deterministic trust, generated through established non-AI tools, to validate AI-produced artifacts. Under this model, AI drives Ada SPARK-based tooling to produce verifiable evidence backed by formal proof, requirements-based test results, structural coverage measurements, and traceability data.
Quentin Ochem, Chief Product Officer at AdaCore, said, “AI can automate many of the steps required to produce high-integrity software, but as with any development project, it requires an instrumented process to build confidence in the correctness of its outcomes. What our tools contribute is deterministic trust. Tools that have been proven over decades to deliver the highest levels of software quality with traditional development methodologies can now be tuned to verify AI-generated artifacts and provide evidence that is trustworthy, reproducible, and actionable. This makes AI-generated artifacts, whether tests, reviews, or even code, usable in high-integrity software.”
The public repository centers around a working traffic light controller model designed for engineering teams to clone and execute independently. The workflow is agent-neutral and has been verified against multiple commercial coding AI solutions to confirm that results remain consistent regardless of the underlying vendor model.
AdaCore, which provides tools for high-integrity development in Ada, SPARK, C, C++, and Rust, offers support for SPARK tooling through its GNAT Pro, SPARK Pro, and GNAT DAS products. By releasing the demonstrator, the company aims to provide practical guidance on adopting AI in a trustworthy manner across both open-source and proprietary projects.
