My #Wikipedia request for comment just closed, finally banning #AI content in articles! "The use of LLMs to generate or rewrite article content is prohibited"

Kudos to all who participated in writing the guideline (especially Kowal2701) and the whole WikiProject AI Cleanup team, this was very much a group effort!

https://en.wikipedia.org/wiki/Wikipedia:Writing_articles_with_large_language_models/RfC

Wikipedia:Writing articles with large language models/RfC - Wikipedia

My genuine hope is that this can spark a broader change. Empower communities on other platforms, and see this become a grassroots movement of users deciding whether AI should be welcome in their communities, and to what extent. On their own terms.

A pushback against the #enshittification and forceful push of AI by so many companies in these last few years.

@quarknova I would say that the battle was lost when Wikipedia allowed big tech to buy access to copyleft content without needing to share alike.

Your new policy simply enforces "fresh meat" for the models, without any requirement for reciprocity back to the commons.

Wikipedians then, are signing up to work for free to feed the models, while people downstream from the models can use their labor entirely for free without giving back.

@yoasif Can I get a citation on there being a way to buy your way out of CC-BY SA? I would agree that there is basically no enforcement mechanism for copyright that the Wikimedia community can lean on other than shame, and that capitalism is shameless.
@yoasif Are you promoting the interesting take that Wikimedia Enterprise selling API access is somehow also relicensing the Wikimedia movement’s CC-BY SA content?
@bd808 yessir.
@yoasif Can we agree that when $COMPANY downloads an xml dump or scrapes the website they are receiving content subjected to the CC-BY SA license if they republish significant excerpts of that content? What evidence can you present that an alternative license covers content that $COMPANY obtains from APIs after signing a contract with Wikimedia Enterprise? Or am I misunderstanding your claims?

@bd808 I think the money that is paid to Wikipedia is an implicit opt-out of share-alike, since the content produced by LLMs are always public domain.

The LLM acts as a copyright removal machine, and big tech is paying Wikipedia to look the other way.

@yoasif Nothing in your concept that any information passed through an LLM becomes public domain is related to usage of a particular API to obtain that information. These companies all already have the content produced by the Wikimedia movement or can have it with trivial effort. This is possible because the movement is about Open Knowledge, knowledge without licensing fees and boarders and resource hoarding. We gave it all away because that was the point from the start.

@bd808 The payment makes a difference.

https://www.avclub.com/wikipedia-ai-partnerships-meta-amazon-microsoft

Wikipedia could have chosen to defend its contributors. Instead, they are taking payment to allow for piracy of contributor works. Given that we know that the big tech LLMs don't respect copyright and derivative works are produced as public domain, the license is clearly being violated, and Wikipedia is being paid for the privilege.

We give it away - but alike, not to be closed. Public domain works can be closed.

Wikipedia intends to make some money from AI scraping its website

Wikipedia intends to make some money from AI scraping its website

AV Club
@yoasif I’m still not understanding how the payment changes anything other than the funds in the Wikimedia Foundation’s balance sheet. Show me the new license terms please. The link you provided is a blog post that looks to be sourced from a Reuters article that in turn was sourced from a press release from Wikimedia Enterprise. The facts I can see there are about a handful of companies having signed up to use APIs where higher rate limits can be purchased. Where is the enclosure?

@bd808 It forecloses the possibility that Wikipedia will defend its contributors.

The license is being wantonly violated, and Wikipedia has co-signed the theft. That is what changes when the funds are exchanged.

Yes, the license has not changed - yet, it clearly has, since Wikipedia is selling an opt-out to NOT defend contributors - even as big tech produces derivative works not protected by the license that contributions were granted under.

What enclosure are you talking about?

@yoasif I’m asking about the enclosure you are claiming. You have I think finally gotten to your thesis: if the WMF is enriched by $COMPANY then they will change their future behavior to protect that enrichment above the rights of the Wikimedia communities. Staying from the start that this is your personal speculation would have been helpful. I disagree with your thesis, but neither of us have a provable position in the near term.

@bd808 I frankly think that they already have - to be paid for access to (newly guaranteed to be human) contributions to feed the models, rather than asking the big tech bots to stop scraping while violating the license is a de facto opt out of the license terms, even as de jure the terms exist.

I am not a lawyer, so it would be handy to consult with one to get their opinion on whether contributors who contribute with the knowledge of these deals are also dual licensing their contributions.

@bd808 That would clearly absolutely be speculation - the view that courts might see a contributor's knowledge of the deals and continued contribution as an acknowledgement that those contributors were given under terms not covered by CC BY-SA -- that is not something that I am claiming, but it seems to me a real possibility to really wrest control of contributions away from the community.

I think the damage already wrought is enough, frankly - Wikipedia has ALREADY granted big LLMs an optout.