The simple fact is that they aren’t able to reproduce their exact training data so no, they aren’t storing it in a highly compressed form.
See: jpeg analogy. You’ve described here lossy compression not something that is categorically different than compression. Perhaps the AI models are VERY lossy. But that doesn’t mean it is original or creative.
But the reality is, we largely do not know how these chatbots work. They are black boxes even to the researchers themselves. That’s just how neural networks are. But the thing I know is they are not themselves creative. All they can do is follow weights to reproduce the things human classifiers evaluated as subjectively “good” over the things they subjectively evaluated as “bad”. All the creativity happened in the training process – the inputs and the testing. All of the apparent creativity outputted is a product of the humans involved in training and testing the model, not the model itself. The actual creative force is somewhere far away.
I see a lot of Dunning Kruger here as well. The fact is that you can generate novel images/texts/whatever with these tools. They may mostly suck but they’re still novel so they can be copyrighted by whoever used these tools to create them.
Even if I grant your premise that their produce is novel – I don’t, that is fundamentally not how they work – the copyright would be held by the bot in that case, not the person who used it.
No more than a person who commissions a painting has copyright for the work. That’s not how creativity, LLMs, nor copyright law works.
I think there’s a lot of Dunning–Kruger here.
See: jpeg analogy. You’ve described here lossy compression not something that is categorically different than compression. Perhaps the AI models are VERY lossy. But that doesn’t mean it is original or creative.
But the reality is, we largely do not know how these chatbots work. They are black boxes even to the researchers themselves. That’s just how neural networks are. But the thing I know is they are not themselves creative. All they can do is follow weights to reproduce the things human classifiers evaluated as subjectively “good” over the things they subjectively evaluated as “bad”. All the creativity happened in the training process – the inputs and the testing. All of the apparent creativity outputted is a product of the humans involved in training and testing the model, not the model itself. The actual creative force is somewhere far away.
I see a lot of Dunning Kruger here as well. The fact is that you can generate novel images/texts/whatever with these tools. They may mostly suck but they’re still novel so they can be copyrighted by whoever used these tools to create them.
Even if I grant your premise that their produce is novel – I don’t, that is fundamentally not how they work – the copyright would be held by the bot in that case, not the person who used it.
No more than a person who commissions a painting has copyright for the work. That’s not how creativity, LLMs, nor copyright law works.
The LLM is a tool. It’s like granting copyright to a paintbrush.
Exactly. Which is how we know that calling what it does inherently creative/novel is absurd and must be wrong. Glad you came around.
Kind of a big jump