Cognitive wheels

Cognitive wheels a system that is computationally powerful but is not psychologically and biologically natural. Artificial intelligence can use cognitive wheels for technological purposes, but cognitive modelling cannot since its goal is to understand the model of the human mind and not just the actions that result from the model.

As criticism of artificial intelligence

The term is sometimes used in criticism of AI, implying that rather than building a system of consciousness similar to the human brain, any AI system is merely a collection of various rules and implications. The theory is that if an AI has acted incorrectly, the system designer must improve the design by adding a provision to deal with the particular case the AI ran into. If this happens repeatedly, the AI does indeed improve, but it could be argued that it does so without a well-designed overarching plan, but rather in reaction to various scenarios. Also, because the system adapts only when the designer adds a cognitive wheel rather than adapting on its own, it is viewed to be flawed and dependent upon the designer.

Response to criticism

Critics of the theory of cognitive wheels counter that it is based on a misunderstanding of the methodological assumptions in the field of AI. They argue that AI models are describable at many different levels - from the most global down to the program code - and the processes of biology and psychology thus cannot be mapped directly on to the AI model. While the areas of implementation detail below the high level of modeling could be described as cognitive wheels, the higher structure of the model is not being accounted for.