More than thirty years have passed since the first drafts of the classical data-information-knowledgewisdom (DIKW) model appeared in the scientific literature. After that - in a society increasingly digital and connected - the DIKW pyramid became popular, featuring variants containing only data, information and knowledge, in a short DIK configuration, or adding levels like intelligence; at the same time that very little critical interpretation and empirical validation were performed by researchers to clarify how these elements are indeed intertwined. This left theoretical implications of the original model unnoticed, not empirically validated, becoming a taken-for-granted explanation that do not fully grasp the chain of knowledge creation, a process that has been carefully studied by researchers. This study reviews and systematically analyzes relevant papers, covering the period of 32 years of research, to pinpoint the main weaknesses of the DIKW model and propose a new one in conformance with the knowledge management literature, and considering the current scenario of artificial intelligence and ethical dilemmas. The resulting model challenges the pyramid as the best way to convey this 'causal process' of knowledge creation to the public, and makes the necessity of empirical studies to be performed in the future clear.