Artificial intelligence (AI) is growing in importance in all societal applications. From inter-sectoral and inter-disciplinary application domains, AI is being increasingly integrated. The road to (results from) AI include a lot of aspects that not only have to comply with regulations, but also have to be tested against deontology and/or feasibility. 
In this project we look at what challenges make AI not directly applicable for certain dynamic processes, and what are its limitations. One of the most (controversial) experience in practice is the fact that data volumes are (very) large, but the information content is limited to offer AI a solution that is also directly actionable. We develop methodologies and algorithms to deal with these limitations in data content/accessibility and allow complementary AI results to combine with traditional methods for the optimal result. The research fits within the framework of modeling and control of dynamic systems.