Wouldn’t this be well handled by suing the person that prompted and distributed the results?

Isn't OpenAI distributing content in its apps?

This is an extremely dangerous precedent that I think you are purposefully trying to put forward.

It's a horrendously bad idea especially for startups to make it apps' faults for how users use their platform. It's only in the benefit of entrenched tech companies to make this precedent.

No, whoever operates the LLM service is liable for unauthorized modification, reproduction and distribution of copyrighted work to users.

Ok, so is it ok if I run the whole thing on my own hardware, and never distribute?

If not, how does that differ from me making an unauthorized pencil drawing of Mario?

It's no different than pirating. No one will care unless you start sharing or try to start a business from it.

Especially true if that child or its mother has a huge market capitalization, large profit margins, highly-paid employees and shareholders eager to reap some more $$.

If the public starts to see LLMs as highly sophisticated copyright laundromats it would most likely hamper further investment & development in that field.

> Especially true if that child or its mother has a huge market capitalization, large profit margins, highly-paid employees and shareholders eager to reap some more $$.

This is the bit I don’t get from the “feed everything to machine” LLM-maximalists. Do they think courts don’t take context into account, do they think all actions happen in a vacuum and that they can just skip along and ignore laws at their pleasure because “tee hee it’s totally definitely fair use bro, I’m totally an academic researcher-pinky promise”.

LLM bros ought to stop and have a think before they poison their own well, assuming they haven’t already done so.

>This is the bit I don’t get from the “feed everything to machine” LLM-maximalists. Do they think courts don’t take context into account, do they think all actions happen in a vacuum and that they can just skip along and ignore laws at their pleasure

An entire generation of unicorn startups believed that (Uber, AirBnB, etc.). We see in the news every day that once you have enough money laws don't apply to you (most things Elon Musk does, the fact that Trump can defy court orders repeatedly and not go to jail, etc.) so yes, this seems entirely plausible.

> Uber and AirBnB

The 2 darling startups that are now facing increasingly less rosy futures?

Airbnb in particular is facing enough backlash that I’d be surprised if it lasts terribly much longer.

Sure, they get away with it for a while, but not forever.

> We see in the news every day that once you have enough money laws don't apply to you

I agree with you here, but I think this is a much broader conversation about capitalism in general which would be getting a bit off-topic for this particular thread, except to say, capitalist forces aren’t above cauterising a limb if it becomes too annoying or intrudes on the other limbs too much. I think the “AI” limb might be overstating its own importance, and I suspect that if it got too up in everyone’s interests re-profit, it would, as an industry, very quickly find itself being neutered. Capital interests would love to get rid of pesky human labour, but if the alternative is too annoying, they’ll have no objections to going back to grinding people through the system again.

AirBnB will get away with it forever. While short term rentals might get banned in a handful of cities, the service now operates worldwide. The stock might be overvalued but if you examine their financials it's simply not plausible to think that failure is imminent.

> The 2 darling startups that are now facing increasingly less rosy futures?

As of this moment uber is worth 120 billion and AirBnB is worth 80 billion.

Yes, they got away with it.

Sure. But if the child has that capability, it doesn't automatically make them a walking copyright violation. "Intelligence", even the current version of AI, entails knowing about stuff, including being able to recite. That doesn't mean intelligence's existence violates copyright. If a person used AI to make a copyright violating work, that's a different story, just like if they used their own innate intelligence to do so.

Taken to its conclusion, liability is then on everyone who decides to publish anything that ChatGPT “tells” them, because it might cross the threshold on plagiarism.

Are the OpenAIs of the world ready to shield their customers from that liability?

If it turns out that using ChatGPT to help you write your resumé opens you up to accusations of plagiarism, or DALL·E to create an image for your website opens you to copyright violation, will you use them?

> Taken to its conclusion, liability is then on everyone who decides to publish anything that ChatGPT “tells” them

Yes. Just like reading anything else on the internet. An LLM is no different from typing "popular cola logo" into Google search and claiming you invented it. If I type "cola logo" into DALL-E and get a replica of Coca-Cola... that doesn't mean I created that logo and can exploit it for commercial purposes.

> Are the OpenAIs of the world ready to shield their customers from that liability?

Why would they? We aren't suing pen manufacturers because someone wrote something libelous using their pen. We aren't busting down the doors of Crayola because little Johnny used the crayons to draw Mario.

OpenAI might not want to shield all their customers from liability, but that is exactly what GitHub have done with Copilot. It's not a hypothetical, it's being done today.

Otherwise it wouldn't get used.

I mean get this great auto complete; if you use it, your code might be AGPLed for all you know, and you're in violation, because you didn't even add a notice.

Would you pay for that?

In a heartbeat. It's time for the old paradigms to die and new ones to be formed.

If ASI can exist I don't believe our the old methods of intellectual fortifications will continue to work in the future. Much like castle walls aren't used to protect against guided missiles.

In that case, the person legally liable for publishing the material is sued for infringement of the work. You don't send someone to jail because they're simply capable of infringing; they have to actually do it, and you have to actually show the specific work whose copyright was infringed upon.

You can also get into the weeds of what's copyright-able (ask Donald Faison about his Poison dance). If you ask for C-3PO and you get C-3PO as he appears in Star Wars promotional material, that seems cut and dry. What if you ask for a "golden robot"? What if you get a robot that looks like C-3PO but with a triangular torso symbol instead of his circular one? What's parody, what's fair use?

Agreed. Fortunately, AI does not have “own works”

You will have to define "own works". How can you measure how "owning" some work is?

AI can’t be a copyright owner. Ergo the violation is on the person using the tool.