Fake News have become a global phenomenon due to its explosive growth, particularly on social media. How to identify fake news is becoming an extremely attractive working domain. The lack of a sound, well-grounded conceptual characterization of what exactly a Fake news is and what are its main features, makes difficult to manage Fake News understanding, identification and creation. In this research we propose that conceptual modeling must play a crucial role to characterize Fake News content in a precise way. Only clearly delimiting what a Fake News is, it will be possible to understand and managing their different perspectives and dimensions, with the final purpose of developing any reliable framework for online Fake News detection as much automated as possible. This paper discusses the effort that should be made towards a precise conceptual model of Fake News and its relation with an XAI approach.