@EUCommission

you already know the correct and only answer.

permission is permission.
right of ownership is right of ownership.
unlawful copying is unlawful copying.
and actions have consequences.

that is the full scope of pertinent information.

therefore if IP is unlawfully copied then the consequences for that apply.

@EUCommission

@EUCommission When that content is paid for by LLM companies is the only correct answer.

@EUCommission I'm not an EU citizen. I may qualify for citizenship in two separate countries because of my ancestry that are member states, I'm a resident seeking citizenship in a country that may soon be an "associate" member, but as it stands, I am not an EU citizen.

I hope you won't dismiss what I'm about to say based on that, but I expect you will at least prioritise the feedback of citizens.

My gut reaction to this was, "Never?" This felt unreasonable given copyright does not forbid access, it just mandates consent of the holder.

Then I thought about why that was my gut reaction and came to the same conclusion.

Not because I think copyright should forbid anyone, I disagree with a lot of how copyright works today. I think the answer is never because they have already stolen countless copyrighted works to train the models already used today. Court cases in multiple jurisdictions have already documented this. They did so mostly using websites explicitly designed to circumvent copyright where using such websites is illegal in most jurisdictions on the planet.

The answer is never because they should be paying exorbitant fines if they aren't already in prison for the breach of copyright they have already committed en masse.

@EUCommission #LLM models should never be trained on content without consent of the copyright holder because otherwise, as already admitted by insiders, this amounts to theft on an unimaginable scale. Not seen since the slave trade and enclosures of common land.

That you are consulting about this is a disgrace and looks like another lobbying driven shifty scam which has become the norm in the EU and elsewhere. Paid lobbying is power for the wealthy and has no place in a democracy.

@EUCommission I will submit a feedback form later but I'm super puzzled why this is even a question at all?

AI companies are not special. They are just like any other for-profit software company, so copyright law applies to them just like it does for everyone else.

@EUCommission I am halfway through the first page of your survey and basically every question assumes that I am okay with AI having more rights than humans. I AM NOT.

This way, your results will be basically useless and just confirm your initial bias.

It’s fine to reply here but fill out this form if you want to have any impact. https://ec.europa.eu/eusurvey/runner/targetedconsultationcopyright

@EUCommission I did my best to fill out the the questionnaire, but it is almost unreadable if you're not a lawyer and from what I did understand it seems very biased. I can't shake the impression that you seem to want to facilitate the whims of corporations instead of serving the interests of the people.

Because why is copyright suddenly an issue when is doesn't suit corporations anymore? Just apply and enforce the same laws that are used to prosecute or fine citizens that torrented a movie.

@EUCommission Only when, when the LLM produces infringing content (content that if produced by a human creator would be considered a close enough duplicate to be infringing), the entity operating the LLM is subject to the same penalties that the human creator would be.

The entity that requested the output is irrelevant. Even if it's the author of the original work asking the LLM to produce a copy of it, it's still infringing (unless the LLM operator can produce a written...

@EUCommission
It should not.

@EUCommission I would accept AI training on copyrighted material if and only if two conditions are met:

1. Humans must be allowed to download the same material for personal use. No more lawsuits for someone who torrents stuff for education or entertainment.

2. Everything that an AI model generates must be marked as such and automatically enters public domain. It has been trained on everyone's data so it should *be* everyone's data.

Essentially we would need to reform copyright altogether.

@EUCommission Only with permission of the copyright holder.

@EUCommission ask tech companies why they dont want employees reading patents. AI is IP laundering.