Those claiming AI training on copyrighted works is “theft” are misunderstanding key aspects of copyright law and AI technology. Copyright protects specific expressions of ideas, not the ideas themselves. When AI systems ingest copyrighted works, they’re extracting general patterns and concepts - the “Bob Dylan-ness” or “Hemingway-ness” - not copying specific text or images.
This process is more akin to how humans learn by reading widely and absorbing styles and techniques, rather than memorizing and reproducing exact passages. The AI discards the original text, keeping only abstract representations in “vector space”. When generating new content, the AI isn’t recreating copyrighted works, but producing new expressions inspired by the concepts it’s learned.
This is fundamentally different from copying a book or song. It’s more like the long-standing artistic tradition of being influenced by others’ work. The law has always recognized that ideas themselves can’t be owned - only particular expressions of them.
Moreover, there’s precedent for this kind of use being considered “transformative” and thus fair use. The Google Books project, which scanned millions of books to create a searchable index, was found to be legal despite protests from authors and publishers. AI training is arguably even more transformative.
While it’s understandable that creators feel uneasy about this new technology, labeling it “theft” is both legally and technically inaccurate. We may need new ways to support and compensate creators in the AI age, but that doesn’t make the current use of copyrighted works for AI training illegal or unethical.
the “Bob Dylan-ness” or “Hemingway-ness”
This is a dumb argument and it’s still wrong. Likeness is protected by copyright laws. See Midler v. Ford.
Fucking Christ I am so sick of people referencing the Google books lawsuit in any discussion about AI
The publishers lost that case because the judge ruled that Google Books was copying a minimal portion of the books, and that Google Books was not competing against the publishers, thus the infringement was ruled as fair use.
AI training does not fall under this umbrella, because it’s using the entirety of the copyrighted work, and the purpose of this infringement is to build a direct competitor to the people and companies whose works were infringed. You may as well talk about OJ Simpson’s criminal trial, it’s about as relevant.
So the issue being, in general to be influenced by someone else’s work you would have typically supported that work… like… literally at all. Purchasing, or even simply discussing and sharing with others who may purchase said material are both worth a lot more than not at all, and directly competing without giving source material, influences, or etc.
If it is on the open internet and visible to anyone with a web browser and you have an adblocker like most people, you are not paying to support that work. That’s what it was trained on.
