Internet Archive (IA) is a familiar name for many, particularly researchers from the Global South who confront on a daily basis paywalls in front of books and scholarly articles they’d like to read. IA is a non-profit organisation that has digitally archived more than 835 billion web pages and 44 million books and texts. It has also archived millions of audio recordings, videos, images, and software programmes. Its archival efforts have played a major role in preserving invaluable global cultural artefacts and has democratised access to them for anyone with an internet connection.

Books form a major part of IA’s archival mission and the extent of access to digitised books is generally determined by the type of user. For example, persons with disabilities can get access to the full text of books that might be still under copyright protection, while others may generally get only a short preview of the book.

The problem experiments

Two of IA’s techno-legal experiments have led to a major copyright tussle with book publishers.

The first is the idea of controlled digital lending (CDL), whereby books are digitised and availed to readers on a 1:1 owned-to-loaned ratio, like traditional libraries. (When a physical library has two physical copies of a particular book, only those two copies of that book can be made available for lending at any point of time by that library.) With the help of digital technologies, CDL extends the same approach to lending books online. The maximum number of books that are available to lend corresponds to the number of books owned by IA or its partner libraries.

The second experiment was the liberalisation of the strict 1:1 lending policy during the COVID-19 pandemic as part of a ‘National Emergency Library’. IA followed this approach for less than three months. It reinstated the lending controls as soon as the books’ publishers initiated legal action citing violation of copyright. The matter came to be titled Hachette Book Group, Inc. and others v. Internet Archive.

The four horses of ‘fair use’

As the ‘National Emergency Library’ project wound down, the core legal issue before both the district court and the appellate court has been the legality of CDL. Since CDL involved making digital copies of the books, and since IA didn’t deny lending those digitised books, the question was whether CDL enjoyed the ‘fair use’ exception under U.S. copyright law. Through a summary judgement in favour of the publishers, the district court took the view that IA did infringe copyright and that CDL didn’t qualify for ‘fair use’.

On an appeal filed by IA, the appellate court reaffirmed the district court’s verdict in favour of the publishers. U.S. courts generally look at four factors as part of their ‘fair use’ analysis: purpose and character of use; nature of the copyrighted work; amount and substantiality of the portion used in relation to the copyrighted work as a whole; and effect of the use upon the potential market for or value of the copyrighted work. The appellate court’s perspective on two of these dimensions is worth reflecting on.

Reading between the lines

Whether the disputed use is ‘transformative’ is an important sub-factor of the ‘purpose and character of use’ condition. It has often had a big influence on the final outcome in ‘fair use’ analysis. On this dimension, drawing on legal precedents, IA argued that their use made book-lending more efficient and enabled use-cases that weren’t possible with print books or physical lending. But the court took the strong position that IA’s actions didn’t constitute ‘transformative’ use as their copies served the same purpose as that of the original works, without meaningfully adding new or different features.

As for the court’s approach vis-à-vis the fourth factor — effect of the use on the potential market for or value of the works: some judges and scholars are of the view that this should be the most important factor in ‘fair use’ analysis. Here, the appellate court held that IA bore the burden of proving there was no market harm for copyright holders. According to the court, while the publishers hadn’t produced any empirical evidence, it was “reasonable and logical” to conclude IA’s digital books worked as a competing substitute for licensed editions of their books. The court also opined that if IA’s practices were to become unrestricted and widespread, they would annihilate publishers’ markets across formats.

Thus, the court discarded the evidence IA had shared regarding the reportedly negligible effect CDL would have had on the sales of copyrighted work and data showing the possibility that CDL could have increased the popularity or sales of books in some cases.

Finding that all four ‘fair use’ factors weighed against IA, the court concluded that IA’s ‘fair use’ argument failed in law in this matter.

The future of CDL

It’s possible to a great extent that the IA’s ‘National Emergency Library’ triggered a panic among the publishers and also predisposed the judges to this outcome. Importantly, however, the long-term consequences of the court’s approach could matter more.

CDL has had a significant effect: it made the book-lending enterprise more efficient in the digital space and era and ensured people around the world could access books in a legitimate manner (without resorting to piracy, e.g.), and without causing substantial economic harm to the authors and the publishers.

On the flip side, if we extrapolate from the court’s logic in this decision, we could argue that even lending by physical libraries falls afoul of the law because book publishers could argue that physical libraries are eating into the sale of their books. However, as experience has taught us, book-purchasing practices can work differently. That is, readers will continue to buy books despite their being available in a library. Copyright laws in different jurisdictions have allowed libraries to let patrons borrow books in view of the broader public benefits, even if it may cause some economic harm to publishers and/or authors.

Importance of real-world data

As the appellate court itself said in its judgement, the publishers didn’t provide any empirical data to substantiate their claim of economic harm. Thankfully, Indian courts have shown the wisdom to seek empirical data from publishers in litigations with similar legal questions. The Delhi university photocopy shop judgement, on the use of copyrighted materials as part of university course packs, is a remarkable one in this regard.

Relying on superfluous logical inferences instead of real-world evidence can distort the balance that ‘fair use’ maintains within the copyright system. If similar litigation arises in India, we should hope Indian courts will differ from the U.S. decision in the IA case and uphold the broader public interest in enabling CDL.

Arul George Scaria is the DPIIT IPR Chair Professor at the National Law School of India University (NLSIU).