Webinar Description
Key Takeaways
- Webinar addresses how to establish trust in AI-generated code for high-integrity software systems using deterministic, non-AI verification tools.
- Features a demonstration of GNAT Foundry: Intersection, AdaCore’s approach to trustworthy AI development.
- Targets software engineers, safety engineers, and technical decision-makers in regulated industries including avionics, automotive, defense, medical, rail, and space.
- Explores the cost implications of AI in high-integrity software and the appropriate role for human review.
About the Webinar
AdaCore hosts this webinar on October 15, 2026, presented by Tony Aiello, Product Manager for SPARK and AI. The session centres on a demonstration of GNAT Foundry: Intersection, a tool designed to address trust challenges in AI-generated code for safety-critical and security-critical applications.
The webinar challenges the conventional approach of relying on human review to validate AI-authored code. AdaCore argues this method does not scale effectively and fails to leverage human strengths. Instead, the session proposes using deterministic trust provided by non-AI tools wherever possible, allowing human experts to concentrate on areas where their judgement is essential.
Deterministic Trust and Tool-Driven Verification
The core concept presented is deterministic trust: using established verification and analysis tools to provide repeatable, verifiable assurance of code quality rather than depending solely on subjective human assessment. AdaCore’s toolchain, including SPARK Pro for formal proof, GNAT Static Analysis for Ada, CodeSonar for C/C++, and the GNAT Dynamic Analysis Suite, supports this approach across multiple programming languages including Ada, C/C++, Rust, and SPARK.
This tool-driven methodology aims to reduce the cost and risk associated with high-integrity software development while maintaining the rigorous standards required in regulated industries.
Topics Covered
- How AI may reshape high-integrity software costs
- How AdaCore tools provide deterministic trust in artifacts produced by AI
- The human’s role in AI development
- Future directions for AdaCore in AI development
Who Should Attend
The webinar is designed for mid-level to senior technical professionals working in high-integrity environments. This includes software engineers, AI and ML engineers, quality assurance and safety engineers, technical leads, and engineering managers. Organisations operating in avionics, automotive, defense, medical, rail, space, and other regulated sectors will find the content relevant to their compliance and safety requirements.
Format
The virtual webinar includes a technical product demonstration followed by discussion and Q&A.