The SEAMLESS project focuses on upgrading system engineering methodologies so that they can ultimately reduce the likelihood and cost of failures of AI-based systems in an unpredictable world.

Problem Context

AI-based systems have become more and more ubiquitous in both industy and society. They are integral to visions of how to address societal challenges, like greener transportation with autonomous vehicles and smart traffic, or within developments like Smart Industry, which increases competitiveness and reduces resource consumption. The introduction of AI in society is, however, hampered by a lack of trust in AI-based systems. This is particularly strong when the likelihood of failures (and their associated costs) are unknown. For this reason, the field of Systems Engineering and Life Cycle Management focuses exactly on this topic. It aims to reduce the likelihood of system failures and their associated costs. An increased excellence in the Systems Engineering of AI-based systems is therefore crucial for the proper integration of AI-based systems in society.

Unfortunately, current Systems Engineering methodologies can neither assure reliable performance nor safe operations of AI-based systems in an unpredictable world. Methodologies of systems engineering need to be upgraded and partially re-invented to be ready for AI-based systems, especially to cover adaptive and self-learning systems.


The SEAMLESS project therefore investigates and develops capabilities on how to engineer and govern systems that incorporate AI functionality, AI components, or AI systems, such that they are trustworthy during their whole lifecycle. Specifically, SEAMLESS will focus on the development of a digital model that supports the verification & validation, continued performance & maintenance, and oversight of AI-based systems. The model will consist of a digital twin and a digital thread that together form the single source of truth of the knowledge on a system. The digital twin will allow for modelling and analysis, while the digital thread will form a link between the digital and physical artifacts, including their changes over the lifecycle. Once developed, it will be possible to implement the digital model as a lifecycle management tool.



  • Joke Welten, Deputy Research Manager Integrated Vehicle Safety, TNO, e-mail: