Understanding the brain, perception, and action paradigm
An agent can be defined as an entity that has the capacity to act. In philosophy, an agent is a being that also possesses desires, beliefs, and intentions. Traditionally, there is an overlap between an “agent” and a conscious entity. A conscious entity should also possess its own internal state that enables it to understand the world according to its internal representation.
An AI agent is defined by its ability to perform an action, but it does not possess desires and intentions (unfortunately, as we discussed in the previous chapter, a model inherits the biases of its training set and thus we can vaguely speak of beliefs). An LLM possesses an internal state, but it is merely a learned representation from the data it was trained with. So no, an AI agent is not a conscious entity. Although the term agent (and others such as representation and internal state) has a different meaning in philosophy, calling an LLM...
