The Advent of Artificial Intelligence (AI) Is Reminiscent of Photography in the 1800S, with a Big Difference

AI and photography share the essence of automating knowledge acquisition

Luca Vettor
3 min readMar 9, 2023
Generated with DALL-E — Prompt: “artificial intelligence is reminiscent of photography”

Comparing AI and photography helps us understand what’s happening with ChatGPT and all similar technologies.

Sure: they are profoundly different due to the significant contrast I’ll describe. Yet, they share the same intention: automating how to represent the world.


Before photography, painters were in charge of visually representing the world.

Centuries of painting techniques have led to more accurate and realistic visuals. Let’s think of perspective, for example. Discovering the perspective and realistically painting the space as our eyes perceive it took centuries.

Until photography, a reliable image of anything required study and effort. In other words, it required an expert: the painter.

Then, it’s become a matter of pushing a button on a camera.

Sure: as a painting can be awful, the same is true for a photo. So, taking a great photo is more than just pushing a button on a camera. It’s about seeing the world and being proficient in using the camera.

Yet, a bad photo requires no effort with a camera, while a bad painting requires considerable effort. That’s the point: automation.

After the advent of photography, have painters disappeared? No, they have not. Have they changed their purpose? Yes, they have. Painters that didn’t change and tried to compete against cameras lost and disappeared, only them.

AI and the big difference

Back to AI, there’s something profoundly different compared to photography: if photography automates getting an image from the real world, AI automates presenting already existing artifacts.

You don’t have to train a camera to take photos — even though you can train it to take better photos than the camera physically could take. Instead, you must train AI to get an answer from it and training grounds on telling the machine what’s right and wrong based on an already existing artifact.

To get an image from an AI, you must provide it with billions of existing images. In other words, AIs can only answer a question if a similar answer exists and is part of its training.

That’s the one big difference between photography and AI: the former has the world as a source of information, while the latest has human artifacts as a source of information.

That means that photography explores the world while AI explores what human beings have already explored of the world. Even when it comes to big data, that massive amount of data is a human standpoint about its source. Collecting data is choosing some of them and excluding others.

In other words, AI is human introspection. While it appears robotic, it’s profoundly human because it embeds in the human judgment of what’s right and wrong.


Will AI help us understand humankind? That may not be the aim. Nevertheless, AI mirrors humankind’s information storage because it reproduces the average way of answering questions.

AI is blurred photography of human knowledge that re-focus when interrogated.

AI is human introspection. While it appears robotic, it’s profoundly human because it embeds in the human judgment of what’s right and wrong.

Compared to photography, is there anything new here? Like photography automates the perspective, AI automates the presentation of the search in the sea of human knowledge.

Regarding automation, the novelty lies not so much in the instrument itself but in the possibilities that it offers: we are exploring them.

If you enjoyed my article and found it helpful, please consider joining Medium through my referral link: you’ll support my writing and get into a sea of knowledge. No extra costs!



Luca Vettor

My 24 years in the IT industry and physics degree flow into my mission: simplify what appears complex.