No.

Yes. 

The issue is not the Elfquest IP. The issue is that any model capable of making an Elfquest movie would require training from data that is not owned by the Pinis. 

And 85% of this crime has already happened. Past tense.

These now Trillion dollar companies are working on the premise that is better to ask for forgiveness rather than fight for permission.

But I veer away from the "No human" element, since a lot of what it regurgitates was made by some very talented humans. It stands blindfolded on their shoulders. Some lessons that may have taken a lifetime now repackaged in a few seconds. The horrific part is depending on the competency of the prompter, these lessons go past unnoticed by both the AI system and also the prompter.

Whatever you think about "prompting" it is also something that is a point of potential human interjection, although it is something that can be incrementally relinquished more and more to AI agents.

Corporations claim it is too difficult to repay the individual artists that have been robbed, mostly because they have no interest in the massive amount of work that would be required in order to find everyone.

So now you are seeing other corporations, mostly from the music space coming after billions here and there for infringment.

It'll take some time, but I'm sure the law will eventually catch-up.

Ethics seem to have been the baby in this particular bathwater. 

As amazingly complex as it seems, it still just a rule following echo chamber. Nowhere even remotely near 'thinking'. It is closer to a gambling app than thinking! In-so-much that it predicts the anticipated result.

The way it is currently set up, you use a prompt and that starts the process. However, if you don't care about the specifics of the outcome, that prompt can be very sparse. 

For example, 1975 Batmobile would indeed give you a quality image.

At the same time, I'm confident they could set it up to go hallucinate by itself and no prompt would be needed.

The frustrating aspect to the whole thing is that if you do care about the specifics and provide a very detailed prompt, the less likely you are to get a satisfactory outcome.

A little will go a long way with AI. And of course you could tell the AI to write your prompt for you anyway. Tell the AI to write you a story. Then tell the AI to break it down into 10 second clips. Then ask for the prompts for those clips. The real legwork is then in policing the temporal inconsistencies between the clips; something that currently  AI is terrible for at the retail level, as the companies hide the seeds that guide the stochastic process.

It's impossible to ethically use AI. The power costs, the impact on the environment, the water use are all unacceptable. This is before AI is trained on plagiarized material. 

The thing that will kill AI is when they finally start charging what prompts actually cost to produce.

Evan S Kurtz wrote:

*******

Your response is a very popular one for generative AI fans, but it's a false equivalence. Human beings don't data-scrape millions of images and books and regurgitate them mechanically. Generative AI isn't conscious: it's a jumped-up photocopier.