art fights ai!

Your Copyright

AI was trained using existing data. Pictures, audio, text, and so on. Many top tier AI companies do not provide proof that they did not violate copyrights.

If it was your work, your copyright was violated.

---

AI

Training the AI means fine tuning a decision algorithm to generate an outcome similar to the training data.

Training an AI wrong results in a bad copy of the input. Training with only one picture or audio snippet results in a copy.

AI could be seen as a mix of a randomizer and an intelligent copy and mix algorithm. The extend might be open for discussion.

But the creation of any copy is usually regulated by law to protect the copyright content. Not in the age of AI. Because proof of violation is usually not available.

---

Discussion

Let us skip any philosophical dimension, because this considerations here are aimed to be fact based and industry related.

I performed a thought experiment as a discussion with the chatbot claude.ai.

This is reasonable, because in terms of intelligence it is difficult to use definitions to distinguish AI from a human.

Thus it represents an efficient summary of the exhaustive training data regarding this topic.

Several scientific publication discuss the issue like "AI language models killed the Turing test: do we even need a replacement?" in Nature, 20 October 2025.

Nevertheless, until AI has the same rights as any human artist, it cannot be considered an independent artist.