Tyng-Ruey Chuang
@trc@social.coop  ·  activity timestamp last month
https://www.internetarchive.eu/protecting-the-past-to-power-the-future-internet-archive-europe-launches-the-our-future-memory-campaign/

> The shift from owning physical materials to licensing digital content has created an unprecedented crisis. License agreements routinely prohibit preservation activities that were once standard practice. Materials that exist only in digital formats often remain locked behind commercial platforms...

> This is not merely a technical problem, it is a fundamental threat to the democratic principle that knowledge should be accessible to all...

More: https://ourfuturememory.org/
Ed Summers
@edsu@social.coop replied  ·  activity timestamp last month

@trc I support this idea. I would personally be more enthusiastic about signing on to an interpretation of Ranganathan's five laws (for web archives) which I've always interpreted as a scientific rather than legal theory en.m.wikipedia.org/wiki/Five_l

Also the access "for advanced research techniques" reads like carte blanche for LLMs, which I do not support.
Tyng-Ruey Chuang
@trc@social.coop replied  ·  activity timestamp last month
@mia @edsu

I feel encouraged as this represents a collective position advanced by the IFLA. Of the 4 rights, #4 on sharing/transferring digital collections caught my eyes as it suggests the group recognize the problems they are facing cannot be solved alone by individual libraries.

I have long admired the consortium funding model practiced by the Open Library of Humanities (https://zenodo.org/records/8398194) and others. Libraries can go beyond acquisition. They can generate knowledge materials too.
Ed Summers
@edsu@social.coop replied  ·  activity timestamp last month

@mia @trc I guess I don't really understand why it needs to be mentioned at all? Research is what people do in libraries, and libraries provide tools to help. What does "advanced blah blah" have to do with it? It seems like another example of how IA's legit concern with controlled digital lending is getting further mired in issues related to LLMs. I wonder who in the organization is doing this? I guess it must be Kahle, but it doesn't seem like a conflation he would make for some reason...
Tyng-Ruey Chuang
@trc@social.coop replied  ·  activity timestamp 4 weeks ago

@edsu @mia Now I read the full statement ourfuturememory.org/full-state I think it is very well said. The first two rights are to collect and preserve which are difficult if the libraries can only subscribe to services then re-lend to the public (Section 1201 Exemptions copyright.gov/1201/2024/ have no teeth here as there is no physical/digital artifacts to acquire; it is now all ToS and paywall). The libraries as such are restricted in the services they can provide to the public, hence rights #3 & #4.
Tyng-Ruey Chuang
@trc@social.coop replied  ·  activity timestamp 4 weeks ago
@edsu @mia Rights #3: "PROVIDE CONTROLLED ACCESS TO DIGITAL MATERIALS for advanced research techniques and to patrons where they are—online." I envision a scenario where the public, say a retired researcher or just anyone, can request access to a customized collection of digital materials not only for personal study but also for text mining, model training, etc. Can public libraries fulfill such requests?

Sadly now we also see BigTech destroying millions of books: https://www.authorsalliance.org/2025/06/24/anthropic-wins-on-fair-use-for-training-its-llms-loses-on-building-a-central-library-of-pirated-books/
Ed Summers
@edsu@social.coop replied  ·  activity timestamp last month

@mia @trc I'm concerned that the perennial good will that exists towards IA is being undermined by IA's stance towards LLMs, which will ultimately lead to their undoing. But, there's still time for them to correct course I think? As you said, this appears to be an EU thing, posing as a universal thing, so perhaps my opinions aren't really relevant here.
Nemo_bis 🌈
@nemobis@mamot.fr replied  ·  activity timestamp last month
@edsu The #InternetArchive has always used a lot of machine learning etc. to extract metadata from its disparate holdings to make them easier to use without having to spend a fortune on traditional cataloguing. Prose generation with LLMs is another matter (but then even the Wikimedia Foundation is getting mired into that nonsense, despite the even more obvious contradictions with its mission).

Getting on the train of this or that legal issue before the courts seems more of a tactical decision.
Ed Summers
@edsu@social.coop replied  ·  activity timestamp last month

@nemobis yes, I think that's right. Libraries (IA included) have always used technology to help provide access to documents. The heart of my concern is the enclosure of the web commons that is being enacted by LLMs driven by huge amounts of capital, and IA's willing participation in that enterprise, when they could just watch from the sidelines, as it were. IA's data must be a very lucrative collection of content for these companies. I worry that the ship has already left the harbor...

1+ more replies (not shown)