Unifying Model for AI Standardization Processes

It is the main claim of this paper that AI standardization processes can be modeled after the basic control structure, thus showing a systematic view of the standards extracted from a rather chaotic milieu of a very diverse standardization conglomerate. The principal idea is to make an analogy between the Controller interactions with external entities, on one hand, compared to respective interactions held at the AI Level, on the other hand. This way, one would have the AI Level interact with Application Domains just like the Controller interfaces with the Plant, the AI Level enabling access through external interfaces, just like the Controller interfaces with a User, and so on, with additional layer allowing Evaluations of the AI Level. Respective examples of AI standards corresponding to the proposed classification, which we call “a unifying model of AI standardization processes” are shown in four respective tables throughout the paper.