Predictive processing is a fresh and new framework in cognitive and computational neuroscience that have been influenced by several disciplines, including artificial intelligence, philosophy, psychology, and so on. One of its main ideas is to see to the brain as a prediction machine: its goal is to anticipate the incoming sensory data (that is predicted) with the actual sensory data (real). The PP framework has been applied to several distinct functions of the brain including action, perception, attention, cognition, etc. Most recently, PP has also been suggested to serve as a framework for consciousness. The main focus in this talk is on whether PP can properly explain consciousness. Consciousness can be characterized by content, level/state, and form. Based on various lines of empirical data, we argue that PP can well account for the content of consciousness. In contrast, PP remains insufficient when it comes to the level/state and especially the form of consciousness including the subjective experience of the contents of consciousness as characterized by various phenomenal features. Hence, we conclude that PP remains limited in explaining the association of content with consciousness (what is known as the “Selection Problem”). Therefore, PP needs to be complemented by a wider and different framework which, as based on the recent temporo-spatial theory of consciousness (TTC), may be spatiotemporal.