Keysight Technologies has entered a research collaboration with the Centre for Assuring Autonomy (CfAA) at the University of York to develop practical methods for validating AI systems in software-defined vehicles. The partnership targets a growing industry problem: as AI takes on greater roles in ADAS, automated driving and SDV functions, auto makers need structured, evidence-based ways to demonstrate safety and reliability across the vehicle lifecycle, in line with tightening regulatory expectations.
The work will concentrate on three areas: building evidence-driven approaches to support implementation of ISO/PAS 8800, developing measurable safety-scoring methodologies rooted in academic research and existing standards, and creating frameworks for producing auditable AI safety evidence. The intent is to support engineering teams in constructing formal AI safety cases and generating the documentation needed to justify confidence in AI-enabled systems, rather than relying on ad hoc validation practices.
The collaboration pairs CfAA's academic work on systems safety, including frameworks already used by transport-sector safety professionals, with Keysight's AI Validation Framework. Simon Burton of the University of York said the aim is to translate AI safety principles into engineering practices usable in safety-critical settings. Keysight's Lukas Klose said the goal is scalable methodologies that help teams generate structured evidence conforming to standards such as ISO/PAS 8800.
Findings from the research are expected to feed into future versions of Keysight's AI Software Integrity Builder, extending its capabilities around safety argumentation, evidence generation and validation for automotive applications.
Source: Automotive Testing Technology International — read the original article