The article reviews some of the relationships between AI and education, emphasizing the metaphors used, the difficulties in finding points of agreement, as well as aspects of the social criticism that is made of AI (e.g. considering that it can be a form of unwanted deviation). AI appears as one more case of technology that comes to improve education, as happened in the past with other less "intelligent" technologies, although now with more arguments in its favor. Secondly, it explores the logical mechanisms (abduction) that are widely used in human reasoning but also in AI, showing how their uncontrolled use can lead to certain conversation algorithms being able to lie, something that teachers would not do. An ethical rather than logical question is raised that deserves to be explored later.