Artificial Intelligence has pervaded contemporary societies in almost every way, namely, as an externalized, fragmented and highly optimized form of rationality. I call this externalized mode of thought “algorithmic rationality” and to the ideology favoring it, “algorithmic reason”. Although algorithmic reason original goal was to facilitate the arising of an highly participative and global Culture, where all citizens would fit in a dynamic and robust democratic society, History has it that AI technology would end being ceased by consumer logic and, more recently, by computational propaganda. I discuss the economical, epistemological and political implementation of algorithmic reason, introducing three cases and arguing that such implementation comprises a cybernetic loop, involving a centralized AI and its users. Commenting upon the ubiquity of such loop, I introduce three limitations for algorithmic reason. The first two are of a computational nature. The third owes its presence to the cybernetic loop, as a steering effect on the human mind is at play, resulting in what I call cultural flattening. This, I conclude, may result in the impoverish of creativity, critical thought and intellectual curiosity in the long run.