Understanding Shadows: Plato's Theory of Forms and AI

Long ago, Plato proposed arguably philosophy’s most famous thought experiment: “The Allegory of The Cave.” It goes as follows:

Imagine a group of people in a cave, sitting with a fire behind them such that it casts shadows onto the wall directly in front of them. A set of figurines is located between the fire and the people to give shape to the shadows. The people are assumed to be unable to turn around and actually see the fire, and all they can see are the shadow figures on the wall in front of them. These people go their whole lives only experiencing these shadows and never gain knowledge of the things actually projecting the shadows. 

Plato uses this thought experiment in Book VII of The Republic to make an interesting point, which is referred to as his “Theory of Forms.” It is essential that, as thinkers, when we understand something, we understand the fundamental essence of that thing, and not merely a manifestation of that thing. 

We observe this in Socrates, interestingly, as he discusses the nature of piety with Euthyphro after meeting him outside an Athenian courthouse. As Socrates is gearing up for his famous Apology, Euthyphro is prosecuting his own father for committing murder. When Socrates asks Euthyphro for his definition of piety, Euthyphro responds, “Doing as I’m doing,” referring to his undertaking of prosecuting his father and obeying justice above family. Socrates hesitates, wondering if prosecuting family is, indeed, pious, but he is far more concerned about the nature of Euthyphro’s answer. It’s an example of what piety could look like, not a definition of its essence. Socrates responds, 

“But just at present I would rather hear from you a more precise answer, which you have not as yet given, my friend, to the question, What is ‘piety’? When asked, you only replied, “Doing as you do.”

Evidently, if we seek true wisdom, definitions that are really only examples seem problematic to our venture. What’s so fascinating is that artificial intelligence has adapted its own version of Plato’s Theory of Forms. 

Let’s say that you ask an AI to generate a picture of a dog. How does it create the requested image? It seems like it would need an understanding of what exactly a dog is, and this is exactly what Plato meant by forms. Through training, AI learns from a wide range of images of dogs, and eventually develop an abstract idea of what a dog is. This process of abstraction seems to be much like what a philosopher does when they encounter Plato’s thought experiment. 

Technology often provides us metaphors for understanding philosophy. What this means is, while Plato did not have AI or even know of its existence, our possession of it allows us to think through these timeless problems of philosophy in comparison to a more tangible concept in our lives, such as AI. Thinking in this sort of “comparison” we can even see where AI falls short of representing Plato’s Theory of Forms, and the deeper question that follows from this thinking:

Unlike a philosopher, who learns through reason, AI learns through repetition of examples. In Plato’s Allegory, this would be like if the people watching the shadows learned what a cat was by seeing a thousand different shadow-versions of a cat cast on the wall. Do they really know what the essence of a cat is if they only see the shadows? Probably not. 

This makes us wonder, “Can a large enough quantity of examples of a thing give us that thing’s form?” Can our experience teach us anything? It’s a worthwhile question to ask.