original melody, chords and lyrics never generated copyright infringement no matter what the composer listened to.

The problem with the music industry is the industry part : it's been generating standardized, and easily mass produced music for the past 20 years. You can't expect the process to not be fully automated eventually ( i've already felt like some pop or rap song could have been for a long time).

You can't expect it not to backfire at some point. And IMHO this is well deserved.

No one says ai companies shouldn't automate what can be automated, but they need to pay for or not use licensed content. Be it art, software, books, anything. If we started stealing cars and selling their wheels we'd be pretty popular too. But we need to pay for what we sell.

An artist « uses » licensed content all the time, just to learn. If we start taxing the learning process (not the teacher, but the learning material itself), then we’re in trouble.

And it’s going to be even worse with multimodal ai and robotics : imagine a robot walking in the streets, looking at cars, and ads, and people, then generate original content based on what it saw on the street (much like a human do). Who’s going to collect for all the licensed content it was exposed to ?

That's the idea - content used for training ais should be licensed. Also there is nothing intelligent about current ai, there is no learning involved, just software ingesting content, there's nothing original about what ai generates. It really shouldn't be compared to humans. Just because a laptop can emit sound it doesn't mean it can sing. Neither can a piece of software aimed at mimicking human brains think or learn.

But then you'd make a distinction in the license based not on the usage (learning), but on the recipient. A human listening to a song and "learn" composition => fine. A machine doing the same => not fine.

That's going to be totally impossible to apply. Imagine a sound processing software than "learned" mixing based on a collection of popular records, and provide you, the composer, with presets => wrong, copyright.

Ok, now imagine an ai that learned how to configure a sound processing software based on records. It only outputs numbers and presets for the software. => wrong, copyright.

Now, imagine an ai that learned those settings, by asking hundreds of human what they like in some kind of A/B testing. The humans will judge based on the songs they've heard in their lives, and the end-result (the settings) will basically be the same as in the previous example. But somehow, this will be fine.

Honestly, i don't see how this can fly.

AI models aren’t databases either though?

Most current "AI" models essentially compress and then store the information, often to the point where training data can be extracted [1][2]

[1]: https://arxiv.org/abs/2012.07805 [2]: https://arxiv.org/abs/2301.13188

I guess but it’s not a traditional database, if anything that can store data is a database then a human should count too

They are, but the main difference is they don't store raw data. They store key features which makes it and even more blatant infringement. Basically, they are made of exactly what copyright protects - characteristics. In on over simplified manner of speaking.

The problem isn’t what they store, but what they produce. a « database of songs » is a problem only if gives you access to the copyrighted content. Wikipedia contains the key characteristics of millions of songs (aka their names and author) and yet nobody complains about it.

Those AI produce original content. The means by which they produce it shouldn’t be relevant.

I agree. Basically, the issue is not that AI can reproduce _ideas_ but that it can execute ideas in the same way as the original, or with parts of the original. It doesn't learn. If it could learn openai would create a new programming language, train it for real, and it would be the greatest programmer. The marketing that is it is "intelligent" and a "black box" that does things on its own is just science fiction. A hallucination on the part of those promoting that type of ai.

I'm not sure you can ground a copyright judgment based on whether something is truly intelligent, or just looks like it.

The means through which a piece of content has been produced is totally irrelevant. The original content has to be distinguishable in the new one for copyright laws to apply. This has been judged over and over when sampling became more and more widespread. Some DJ sampled the original songs blatantly and thus had to pay the original authors. Others applied so many sound transformations that it was basically impossible to identify and thus copyright was basically impossible to apply.

This is not true, it clearly does learn

“If it could learn openai would create a new programming language, train it for real, and it would be the greatest programmer.” I don’t understand what you’re saying here. It can invent new languages, but they would those be better than existing ones? It’s less intelligent than a human (though still intelligent) so the language would be worse than existing ones

I get your point, but creative people are in fact like databases. They gather as much input as possible, and synthesize output based on that and a few mutations.