The upcoming growth of the electric vehicles market to accompany profound societal and ecological mutations requires a large-scale deployment of electric charging points. The first milestone of 100,000 public charging points having almost been reached, the next objective of the French government is 500,000 charging points by 2027. This growth incurs new impactful cyber physical security risks for power grid, transportation infrastructures, customers, and operating companies. Therefore, a need to detect and identify these risks is crucial to offer secure and trustworthy charging units for electric vehicles. Our project aims at designing robust data-driven AI algorithms to detect abnormal events and behaviors on charging points. Ultimately, the elaborated algorithms, accompanied with cyber behavioral analysis module and potential attacks’ mitigation actions, will be deployed in production on real charging points.