Table of contents

AI-generated Art and AI-generated code are treated differently

AI-generated Art and AI-generated code are treated differently

Today, while doom scrolling on Zuckerberg’s social network, the algorithm recommended me an image of Sakura Card Captors in the style of the Spanish painter Remedios Varo.

Image of Sakura Card Captors in the style of Remedios Varo created with AI
Image of Sakura Card Captors in the style of Remedios Varo created with AI

When reviewing the comments -I don’t even know what’s in it for me- I noticed that the ones with more likes expressed a strong disdain for artificial intelligence. The accusations were the usual ones: that it was trained using artists’ works without permission, that it plagiarizes styles, and at the same time, that it is despicable simply for not being human.

But if someone talks about how vibe coded a small app, the response from users is not so controversial, or the author of the app is attacked instead of the AI.

Artists hate IA, devs are forced to accept it.

I’ve always been struck by how the reaction to artificial intelligence varies by professional group. The artists have it very clear: they are completely against it. On the other hand, software developers and people working in offices are asked to adopt artificial intelligence as just another tool.

Those who don’t are labeled fatalists or even luddites, and are urged to accept and embrace, stoically, the inevitable future offered by these technologies.

Are artists right about AI?

The truth is that this issue is quite complicated.

Since the start of the hypothetical AI bubble , artists accuse artificial intelligence of plagiarism, as if it were nothing more than a database of modified images that are superficially altered every time an image is requested. But the reality is that AI doesn’t work like that.

AI does not plagiarize artists, but it is trained without their permission.

AI is able to abstract patterns from its training data set, images in this case. Once it abstracts those patterns or archetypes, it uses randomness to generate results that fit those patterns.

This is quite similar to how a human artist works: he collects stimuli from other artists, personal experiences, books, music… and mixes it all together to create something new. Something that, although original, preserves certain references or characteristics of the works that inspired it.

All of this without being conscious, Searle would have said.

The contrast in perception between code developed with AI and images generated with AI.

In the case of code generation using AI, the process is practically the same. The AI detects patterns and then generates new, different code, but based on the archetypes that the LLMs have “learned”.

However, this is where an interesting contrast appears: while image generation is visual and causes dissatisfaction when reflected on the screen, AI-generated code goes completely unnoticed. **It is not seen.

There is no public backlash, no group protesting the generated code. The code does not arouse passions, nor anger about stolen code, like when Devin AI promised to replace programmers perhaps because the visual stimulus is much more intense than the sensation of knowing that an app runs on AI code.

Even when we look at the crystallization of code, in the form of pixels on a web page, we are reacting to the image of the final product, not the code itself.

Moreover, it is virtually impossible to distinguish a web page made by an AI from an incompetent developer. Like the one who developed this website.

Why do code and images made by AIs elicit such different responses?

I find this contrast fascinating. It’s not that one is right or wrong, I’m simply struck by how AI-generated artistic creation can be dismissed as visibly and emotionally significant, while AI-generated code is either ignored or accepted without resistance and considered an expected advancement of technological development.

Perhaps this is because art is a direct form of visual and emotional communication. And being machine-generated, it can be perceived as an inferior form of expression. In contrast, code, by its functional and abstract nature, lacks that emotional charge and goes completely unnoticed.

Humans evolved using sight as a survival tool, code as an abstraction of processes is very recent and was not part of the evolutionary pressures of our species nor of the material reality beyond millions of ones and zeros on a silicon board.

I wonder what will happen to AI in other areas?

What will happen in other areas. For example: will medical diagnoses made by AI be rejected by patients, just because they don’t come from a human? Is human treatment as important when diagnosing a disease or is it limited to patient care? will people be more objective and only look at the percentage of accuracy and effectiveness?

The answer will probably depend on how visible the hand of the machine is… and how much we care about what we can’t see.

Eduardo Zepeda
Web developer and GNU/Linux preacher. Maturity over novelty, better done than perfect. I used to believe in the goodness of cryptocurrencies outside of monetary speculation.
Read more