Artificial means something that is made. It comes from the same Latin root where we get our words like "art" and "artifice". The key question of Artificial Intelligence is whether it is possible to make an intelligence. In fact, the term Artificial here is almost redundant in the field of computer science, as nothing in computer science is not artificial. We don't call software artificial software, and we don't call RAM artificial memory, even though it is more closely analogous to memory than artificial intelligence is to intelligence in its current state.

Intelligence is the ability to obtain knowledge, understand it, and use it. So far, if we take the definitions for Artificial and Intelligence and lay them out, Artificial Intelligence literally means something that is made and can learn, understand, and use knowledge. Notice that I've inserted an implied object into the definition. If I had not, the definition would have been something that is made which represents the very idea of the ability to learn, understand, and use knowledge. It must be the first definition, since it makes no sense for humans to claim to be able to make an artifice of the very concept of intelligence. In that sense, the term Artificial Intelligence is actually a shortened form of the term Artificially Intelligent System, but I prefer to call it an Intelligent Artifice.

Why is it called Machine Learning and not Artificial Learning? Because "Learning", as a gerund of "to learn", means something like "what is referred to when one learns in the general sense; the action, state, or condition that involves the acquisition of new knowledge and experience." "Artificial" means something that is created, and so while it is grammatically correct to place an adjective in front of a gerund, it is not always correctly done so depending on the nouns and adjectives involved. Can you "create" the "act of learning", or can the "act of learning" be "something that is made"? A more correct adjective and gerund combination would be "furious running".

Machines are constructs, whether abstract or physical, which transform energy into another form. Machines are essentially physical translators. Moreover, there was always a distinction between mechanical arts and liberal arts, where the liberal arts involved the letters and philosophy studied by the free classes and the mechanical arts involved the repetitive and physical trades practiced by the slave classes. 

In this sense, it starts to become clear that "machine learning" will not be able to produce a "liberal intelligence", as machines are constricted to the transformation of energy and matter. Machine Learning is more closely defined as Mechanical Analysis, where a machine will analyze and categorize data.

Artificial Intelligence typically involves analyzing and categorizing large datasets and developing an algorithm based on the analysis that is best suited for performing some task. For example, AI text generators will analyze a lot of text written in a similar style, and then be able to produce similar text given some kind of prompt.

AI is better able to solve these kinds of problems than traditional programming, since the algorithms involved would be both very complex and difficult for programmers to create. An AI/ML stack, on the other hand, is able to design its own algorithms to predict what the solution would be. The underlying algorithms are usually unknown to the programmer.

The first step to reframing Artificial Intelligence is to replace the term Artificial Intelligence with what it is actually doing, leaving "Algorithmic Imitation" as the replacement. Algorithmic Imitation is a much better representation of what certain applications of Artificial Intelligence are capable of. An AI text generator is really doing nothing more than imitating a large dataset.

However, even though we still have AI as the acronym, it's not an entirely accurate representation of what is happening. AI's strength is that the underlying algorithms are largely unknown to the programmer. So, in that case, I will modify the new term further to be "Algorithmically Ambiguous Imitation".

The final piece of the puzzle is describing the nature of the system, which is mechanical, since everything about this process is mechanical in nature. This leaves us with the final term, Algorithmically Ambiguous Mechanical Imitation.

AI/ML is really just Algorithmically Ambiguous Mechanical Imitation.