Ideal organization theory describes a classification system that structures the interaction of self-determining actors. It states that the actors need to be formally structured only to the extent that the internal interaction is finite, in other words there can't be too many cooks in the kitchen.

This is what oligopoly really means, a small set of actors, which each actor being able to understand the position of the other actors in the set and their behavior. If the set is too large by definition an actor cannot use the other actors behavior as part of their decision process.

Membership in the oligopoly is subject to terms and conditions, so an actor will only remain as part of that set in-so-much their actions are accepted by the definition of the structure. Even if the member actor is not intelligent, their self-determining actions (be it by randomness or an algorithm) will determine whether their membership is valid.

These organized sets then need to be able to interact within a free marketplace to participate in processes that are non-finite, that is the nature of the interaction makes oligopical interaction inefficient.

When these strategies are combined the resultant marketplace is optimal, that is it is the best combination of efficient and fair Arthur Okun's Equality and Efficiency: The Big Tradeoff for a great discussion of what optimal means in this context. He argues that Efficiency and Equality are on a Laffer Curve.

The resultant marketplace will be artificially intelligent to an outside observer as the structure results in information flow that would elicit a more intelligent response than any one actor or set of actors. Hence artificial intelligence simply emerges from highly optimal organizational structures.