AdaCore Released GNAT Foundry Intersection Tool
The open-source demonstrator offers a framework to verify AI-generated software updates through formal proof.
Updated on Sept. 25, 2026 in Artificial Intelligence
Live Poll
Do you trust AI-generated code to be used in safety-critical systems?
AdaCore has launched the GNAT Foundry: Intersection project to help developers verify code generated by AI agents. The tool uses SPARK to confirm the absence of run-time errors in software updates.
Why it matters
AI coding agents can generate and modify software artifacts faster than engineers can manually review them. This demonstrator provides a workflow to generate independently checkable evidence for these automated changes.
The demonstrator requires 10 GB of free storage and is optimized for Ubuntu 24.04 on x86_64 or AArch64 systems. It successfully runs a complex four-way traffic-light controller scenario that includes pedestrian crossings.
The players
AdaCore
AdaCore is a software development company that specializes in providing tools for mission-critical and high-reliability systems.
The details
The workflow employs deterministic tools to propagate updates through requirements, architecture, implementation, and tests after an AI agent processes a change request. While the project proves the absence of run-time errors using SPARK, it is not currently certified to any safety standard.
Timeline
September 25, 2026: AdaCore officially released the GNAT Foundry: Intersection project.
The Tech Race
This release marks an attempt to bridge the gap between rapid AI code generation and the rigorous safety requirements of industrial software engineering. It positions AdaCore as a developer of trust-based tools in a sector increasingly crowded by unverified generative AI models.
Developers interested in experimenting with the tool can access the repository on GitHub to test their own AI-generated workflows. However, users should note that the software is not intended for deployment on public roads and lacks official safety certification.
The takeaway
Reliable software development requires automated verification tools to keep pace with the speed of artificial intelligence. Engineers should prioritize deterministic checks over manual reviews when integrating AI agents into high-reliability pipelines.
Further reading
Learn more about the latest innovations in Artificial Intelligence to understand how industry standards are evolving.
Source note: This article includes information reported by eeNews Europe.
Live Poll
Do you trust AI-generated code to be used in safety-critical systems?
