A Review and Strategic Approach for the Transition towards Third-Wave Trustworthy and Explainable AI in Connected, Cooperative and Automated Mobility (CCAM)

Traditional "2nd Wave (2WAI)" AI algorithms are highly specialized in narrowly-defined tasks in the transportation and mobility domain, however, quite hardly explainable and unable to cover the needs of unbiased and trusted decision-making in the CCAM domain. There is a strong need to accelerate the shift towards human-like "3rd Wave AI (3WAI)" to overcome the challenges of road transport to make autonomous vehicles and driving safe, cleaner and more efficient. This paper aims to shed light on ambitious pathways toward more trustworthy and explainable AI by presenting a strategic approach envisioning the requirements of 5 main audience profiles in the CCAM context (developers, decision-makers, regulators, end-users, and CCAM Service providers). The presented approach envisages a 6-dimensional concept that brings features of i) Transparent and explainable; ii) Fair and impartial; iii) Responsible and accountable; iv) Robust and Reliable; v) Respectful of Privacy; vi) Safe and Secure, transportation. To achieve the transition towards 3WAI a model-based n-sprint V methodology is introduced that is enriched with continuous validation by putting ethics at the centre of all potential innovations. The paper presents a review of the recently funded European projects within the concept of AI-powered CCAM.