The Role of the Academic Library

In many ways, academic libraries are well-positioned to see the effects of publisher consolidation firsthand, as they remain the primary purchasers of scholarly content. Since the 1990’s academic libraries have largely transitioned their purchasing of academic content to “big deals”, under which publishers bundle together hundreds or even thousands of journal titles into packages. When this model emerged it offered substantial value for libraries as the per-title cost of journals went down. “Following this new model, size of serial collections in academic libraries increased almost fivefold from 1986 to 2011″.^[1] While this new purchasing model allowed libraries to rapidly increase their collections, it also resulted in increases in collection expenditures and decreases in the value of these investments, as researchers and students still largely read and cited only a fraction of the titles held in library collections.

Transitioning Away from the Big Deal

In more recent years, the thinking around the value of the big deal has slowly changed as libraries have come to terms with the fact that it is no longer possible, nor necessary, to keep up with the ever increasing level of scholarly output. For decades now, publishers have been increasing the costs of journal subscription deals at rates that outpace inflation. As a result, subscriptions become more expensive every year. This has resulted in what is often referred to as the “serials crisis”, whereby academic libraries are increasingly unable to keep up with the rising cost of subscriptions.

In the last few years frustration over the rising cost of subscriptions, paired with the growing availability of open access (OA) research, has led many libraries to either cancel their big deal subscriptions or work with publishers to draft new types of licenses that incorporate OA principles and offset costs for OA publishing. In a 2020 survey of academic library directors, half indicated that they will likely cancel a journal package in the next five years.^[2]

While big deal cancellations are still news-worthy occurrences, there is good evidence to suggest that they will become more commonplace in the future. Schools that have yet to make the difficult decision to cancel are looking to trailblazers like the University of California, which recently cancelled its subscription with Elsevier after a long and ultimately unsuccessful renegotiation.^[3] The UC system was able to marshal faculty support for their cause, but many schools are still unsure if their faculty and students will support a decision to cancel a large subscription package.

Although UC eventually came to an agreement with Elsevier in 2021, other universities and purchasing consortia watched their negotiation process carefully and have begun to fight for more progressive terms in their licenses.^[4] For example, in 2021 the Canadian Research Knowledge Network negotiated a new Elsevier license on behalf of its member institutions in Canada that includes reduced costs for OA publishing and other progressive elements.^[5]

Scenario – Breaking up with the ‘big deal’ Let’s consider this scenario: The library has just informed your department that due to budget constraints they plan on cancelling a large subscription package that includes many of the journals you regularly read and cite. The library will be purchasing individual subscriptions to a much smaller number of journals and while a few of the key journals you rely on are on the list, others are not. What would be your initial response to the library’s decision? What should the library do and who should they consult before finalizing their new approach?

The Next Big Deal – AI Training Collections

While libraries are still contending with the legacy of subscription packages and publisher “big deals”, a new frontier for licensing is beginning to emerge around use of collections for GenAI training and text and data mining. Many standard licensing agreements between libraries and content providers explicitly forbid using these collections for textual analysis or AI training. Consequently, researchers are often unable to access library resources in the ways they require. In the long term, libraries are hopeful that AI regulation and policy will help clarify that using copyright-protected works for machine learning is a protected right, but in the meantime, libraries are trying to renegotiate their licensing agreements to permit AI training, or are purchasing special collections that are built for such uses. This inevitably puts new financial burdens on libraries and creates additional hurdles for students and faculty. Additionally, it raises the issue of bias, as researchers may be limited to using inferior data and or tools because of licensing restrictions.^[6]

Libraries are just beginning to see the commodification of collections for AI purposes and many of even the largest publishers have yet to settle on a model for AI licensing. ITHAKA S&R has begun to track generative AI licensing agreements, and have noted that “Thus far a standard set of terms or overall model from which to build these deals has yet to emerge.”^[7]

Dig Deeper To learn more about the cancellation of big deals, review: SPARC’s Big Deal Cancellation Tracker Anderson, Rick. (2017). When the Wolf Finally Arrives: Big Deal Cancellations in North American Libraries (comments on this post highlight the ongoing debate).