2 Attachment(s)
Anna’s Archive Urges AI Copyright Overhaul to Protect National Security
One of the internet's leading shadow libraries is calling for an overhaul of copyright law. Anna's Archive believes that if countries want to stay relevant in the AI race, change is a necessity. While it's unlikely that the trove at Anna's Archive will ever be legally accessible by consumers, the site argues that access for AI companies is paramount. "It's a matter of national security," Anna's Archivist says.
booksThe Artificial Intelligence boom promises unparalleled progress which, in theory, can change the fabric of society.
As startups and established tech giants explore their options, data has become the new oil in the race to develop AI. Archives of text, in particular, are key to training Large Language Models (LLMs).
Early on, many tech giants used shadow libraries to train their models. Rightsholders objected to this unauthorized use in various courts, transforming these repositories into ‘forbidden fruit’ in the U.S. and elsewhere. Such restrictions aren’t necessarily observed by all, creating an uneven playing field where developers in other countries gain an advantage.
Shadow libraries already clash with the law, but copyright holders want to make sure that governments don’t make exceptions for AI training. Getting all countries onboard could be a real challenge.
According to Anna’s Archive, one of the internet’s leading shadow libraries, there is a better solution. This is not so much a matter of choice, the site says, but a requirement for national security.
‘Overhaul Copyright Law to Protect National Security’
Shortly after we published an article on the AI “forbidden fruit” conundrum and potential consequences, site operator Anna Archivist published a call to action, stressing the need for a strong and urgent response.
“If the West wants to stay ahead in the race of LLMs, and ultimately, AGI, it needs to reconsider its position on copyright, and soon,” it reads.
Anna’s Archive is not just a passive observer in the artificial intelligence arena. Prominent AI companies including DeepSeek have used its library of books and articles to train their AI models.
The Race to Artificial General Intelligence
Anna Archivist says high-speed access to data has already been provided to 30 companies, most of which are based in China. These include LLM companies and data brokers.
While some U.S. and Western companies have become more cautious about copyright, others are forging ahead. That has the potential to create problems for the West, the shadow library’s operator says.
“Whether you agree with us or not on our moral case, this is now becoming a case of economics, and even of national security. All power blocs are building artificial super-scientists, super-hackers, and super-militaries. Freedom of information is becoming a matter of survival for these countries — even a matter of national security,” Anna Archivist writes.
The future of AI development won’t stand or fall based on access to a single shadow library. However, it’s undisputed that vast amounts of data have proven to be extremely helpful in getting LLM models to where they are now.
Concrete Copyright Proposals
Since copyright holders label Anna’s Archive a piracy haven, it comes as no surprise that the site is an advocate for copyright reform. Interestingly, however, the site’s proposals are not as extreme as one might expect.
“Our first recommendation is straightforward: shorten the copyright term. In the US, copyright is granted for 70 years after the author’s death. This is absurd. We can bring this in line with patents, which are granted for 20 years after filing,” Anna’s Archivist writes.
A reduced copyright term wouldn’t legalize Anna’s Archive, but it would elevate many more works to the public domain. At the same time, however, there will still be an advantage for those AI companies that don’t have to worry about copyright terms at all.
This leads to the second proposal. Anna’s Archive believes that countries should have copyright exceptions, or carve-outs, for the “mass-preservation and dissemination of texts.”
Exceptions shouldn’t necessarily apply to consumers, but LLM companies, libraries, and archives should be covered.
Diametrical Opposites
Especially when considering the source, these suggestions may seem radical. Yet text and data mining (TDM) exceptions already exist in current copyright law, and in countries such as Japan, carve-outs for AI have already been made.
Not surprisingly, these AI exceptions are a concern for rightsholders. A few days ago, the Digital Creators Coalition (DGA) informed the US Trade Representative that it’s against all AI-related copyright exceptions, period.
“We strongly oppose broad copyright exceptions for AI, including with respect to TDM, and reaffirm our position that its proponents have failed to demonstrate the need for such an exception,” DGA informed the USTR.
“However, some countries are actively considering, or have already adopted, TDM exceptions that fundamentally weaken copyright protection in favor of promoting AI at the expense of the American creative sector.”
This is the diametrical opposite of Anna’s Archive’s position, and likely against the wishes of many AI developers too.
While a shadow library’s calls for copyright reform seem unlikely to be considered, the AI copyright discussion isn’t over. We expect to see similar talking points in the future, also from other sources.
Anna’s Archive maintains that it will continue to operate, regardless of the legal implications.
“As for Anna’s Archive — we will continue our underground work rooted in moral conviction. Yet our greatest wish is to enter the light, and amplify our impact legally. Please reform copyright,” Anna Archivist concludes.
