Estimating the state of charge (SOC) of lithium-ion batteries (LIB) is a critical task that has also become highly desirable as electrified vehicles become more widely used. It is necessary to improve the accuracy of the battery SOC estimation, to develop more efficient, reliable and affordable electrified vehicles. Due to the non-linear behavior of these batteries, an accurate estimation of SOC is still challenging. For this reason, thanks to the greater availability of battery data and advances in artificial intelligence (AI), traditional theory-based methods are often replaced by data-driven approaches. In particular, recurrent neural networks (RNN) should be a promising method to be exploited since they can capture dependencies in time and predict SOC without a battery model.