Automotive manufacturers face a growing technical bottleneck: proving that increasingly complex artificial intelligence remains safe and reliable throughout a vehicle's entire lifecycle. To address this, Keysight Technologies and the Centre for Assuring Autonomy (CfAA) at the University of York have launched a research collaboration focused on software-defined vehicles (SDVs). The partnership aims to bridge the gap between theoretical AI safety research and practical engineering, developing methodologies to generate the auditable evidence required to meet emerging regulatory and industry safety standards.
Bridging AI Research and Automotive Engineering
The collaboration focuses on creating evidence-driven approaches to support the implementation of ISO/PAS 8800 requirements. As AI becomes a core component of advanced driver assistance systems (ADAS) and automated driving functions, the industry is under pressure to demonstrate that these systems operate as intended. Keysight and the CfAA intend to develop measurable safety-scoring methodologies grounded in academic research and industry standards. This work is designed to help engineering teams build structured AI safety cases, providing a framework for generating the documentation necessary to justify the safety of AI-enabled systems. By focusing on practical, scalable validation, the initiative seeks to reduce development risk for automakers and suppliers. The research is expected to inform the future development of Keysight's AI Software Integrity Builder, specifically enhancing its ability to support AI safety arguments, evidence generation, and validation activities that align with current automotive industry expectations and safety-critical environments.
Validating Software-Defined Vehicle Intelligence
The partnership leverages the University of York’s academic expertise in safety engineering and complex system assurance alongside Keysight's existing AI validation techniques. A primary goal is to create practical frameworks that allow for the generation of auditable AI safety evidence. This is particularly critical as vehicles transition toward software-defined architectures, where functionality is increasingly managed through code rather than hardware. The initiative explores how to translate high-level AI safety principles into concrete engineering practices. By combining the CfAA’s research—which includes frameworks already used by transport sector professionals—with Keysight's holistic AI Validation Framework, the collaboration aims to provide automotive organizations with the tools needed to build confidence in AI-enabled systems. This approach targets the entire product lifecycle, ensuring that safety is not just a development-phase checkbox but a continuous requirement. The resulting methodologies are intended to improve the efficiency with which engineering teams can demonstrate justified AI safety, ultimately supporting the faster and safer deployment of intelligent automotive technologies.
Key Takeaways
- Keysight and the University of York's CfAA are collaborating to develop practical methods for validating AI in software-defined vehicles (SDVs).
- The research aims to support the implementation of ISO/PAS 8800 requirements through measurable safety-scoring methodologies.
- Findings from this collaboration are expected to inform the development of Keysight's AI Software Integrity Builder.
TechInsyte's Take
In our view, this collaboration signals a critical shift from theoretical AI safety to the rigorous, auditable engineering required for mass-market automotive deployment. As software-defined vehicles become the industry standard, the ability to provide "justified confidence" through structured evidence is no longer optional; it is a regulatory necessity. By aligning research with ISO/PAS 8800, Keysight is positioning itself to provide the essential verification layer for the next generation of ADAS and automated driving. This move suggests that the industry's primary hurdle is no longer just building intelligent AI, but rather building the standardized, repeatable validation frameworks required to prove that intelligence is safe for public roads.
Questions & Answers
How does this collaboration address the regulatory challenges of AI in automotive systems?
The partnership focuses on creating evidence-driven approaches to meet ISO/PAS 8800 requirements, helping manufacturers generate the auditable safety evidence and structured safety cases required by emerging industry standards.
What specific technical tools will be influenced by this research?
The research is expected to inform the future development of Keysight's AI Software Integrity Builder, specifically enhancing its capabilities regarding AI safety arguments, evidence generation, and validation activities.
What is the primary goal regarding the development lifecycle of software-defined vehicles?
The initiative aims to help engineering teams demonstrate justified AI safety more efficiently throughout the entire product lifecycle, rather than just during the initial development phase.
Why is the University of York's involvement significant for this project?
The University of York, through the CfAA, provides internationally recognized expertise in safety engineering and the assurance of complex systems, offering frameworks that are already utilized by safety professionals in the transport sector.
Source: Businesswire
