AI Certification, Fairness and Regulations
Authors: Bernhard Nessler (Software Competence Center Hagenberg) , Michal Lewandowski (Software Competence Center Hagenberg) , Simon Schmid (Software Competence Center Hagenberg) , Gregor Aichinger (Software Competence Center Hagenberg (SCCH)) , Iana Kazeeva (Software Competence Center Hagenberg) , Rania Wazir (Software Competence Center Hagenberg (SCCH))
The Certification of Artificial Intelligence Systems (CERT AI) workshop, held
under the title AI Certification, Fairness and Regulations at the Austrian Symposium
on Artificial Intelligence, Robotics and Vision (AIRoV) 2026, provides
a forum for discussing how artificial intelligence (AI) systems can be evaluated,
documented, and monitored in ways that support certification and trustworthy deployment.
Its central concern is the operationalisation of certification: translating
legal, ethical, and organisational requirements into measurable, testable, and
statistically valid properties of AI systems. This editorial introduces the motivation,
scope, and programme of the workshop. We argue that certification cannot
be reduced to a final compliance check, but must be understood as a lifecycleoriented
process connecting intended purpose, application-domain definition, risk
management, technical evidence, fairness assessment, privacy protection, explainability,
robustness, standardisation, conformity assessment, and post-market monitoring.
The workshop contributions illustrate this breadth, ranging from statistical
application-domain modelling and General Data Protection Regulation (GDPR)-
compliant machine learning to multi-user conversational agents, safety architectures
for autonomous surface vessels, explainable model selection, anthropomorphic
terminology, and the ongoing standardisation process in the European Committee
for Standardization and European Committee for Electrotechnical Standardization
(CEN/CENELEC). Together, these perspectives point toward a practical
research agenda for functional trustworthiness: AI systems should be certified
not only by what they claim to do, but by what can be verified about their behaviour
in the contexts in which they are deployed.
Keywords:
How to Cite: Nessler, B. , Lewandowski, M. , Schmid, S. , Aichinger, G. , Kazeeva, I. & Wazir, R. (2026) “Editorial: AI Certification, Fairness and Regulations (CERT AI 2026)”, Proceedings of the Austrian Symposium on AI, Robotics, and Vision. 3(1).