Motherboard, an online technology magazine owned by VICE recently published a story about Elsevier’s purported practice of branding their PDFs with a unique digital fingerprint. This means that if you download an academic article from one of Elsevier’s more than 2,800 active journals, it may have a unique identifier embedded in the PDF’s metadata. Identifiers are long character strings; currently we do not know what information they store. Jonny Saunders, a US-based PhD student, broke the story on Twitter.
The question that has arisen since, is why is Elsevier doing this? The publisher said to VICE that “fingerprinting in PDFs allows us to identify potential sources of threats so we can inform our customers for them to act upon. This approach is commonly used across the academic publishing industry.” When prompted for clarification, Elsevier claimed that one use case was to help institutions combat ransomware; ransomware is malicious software used to gain access to an IT-system before holding it to ransom. It is unclear how fingerprinting would prevent this, but what it can do is help Elsevier establish where and when a PDF was downloaded as they scan the internet for illegal sharing of their content. This is particularly relevant in cases where university logins have been compromised and used to harvest PDFs illegally. If those PDFs are then shared on a pirating site like Sci-Hub, Elsevier has a way of identifying and closing leaks.
Elsevier stated to VICE that the fingerprints do not capture personal identifying information or metadata, which is important from a data privacy perspective. Saunders is quoted challenging both the ransomware and privacy claims. As such, we can see the VICE article as a discursive skirmish over the identifiers’ meaning and purpose. I argue that we can interpret the identifiers’ potential meanings and purposes as partial materialisation of a process called assetisation.
The asset connection
I currently work on the ESRC funded project Universities and Unicorns at Lancaster University with Principal Investigator Dr Janja Komljenovic. In the project, we examine the relationship between education technology (edtech) and assetisation. Assetisation is the process of turning things like PDFs or student data into assets. Assets are things that the asset owner can use to generate future income without selling the thing itself, but simply by providing access to it in exchange for a fee, also known as rent. In this way, an asset is very different to a commodity where exchange relations are characterised by a change in ownership.
Rent relations between Elsevier and universities are stipulated in a licencing contract. A contract by itself, however, is not sufficient to maintain the asset state of digital journal articles and books. For example, if Elsevier’s PDFs are accessible through illicit means, then that will compromise the publisher’s ability to generate licencing fees. Fingerprinting is one of many potential ways to assetise a back catalogue of digital content because it gives the company further control over who can access it.
It is understandable that Elsevier aims to improve its ability to spot leaks in its system, support institutions in the fight against ransom ware, and decelerate systematic capture and dissemination of licenced material. However, I am unconvinced that fingerprints and internet takedown programmes are effective ways of achieving such objectives: deleting embedded metadata in PDFs is easy and shutting down pirating sites through the courts is difficult. It would be more effective to spot abnormal activity in real time through anomaly detection algorithms, which could pick up on unusual download patterns. This suggest that fingerprints may also serve other purposes, or have the potential to do so, at a time when Elsevier’s reputation in the academy is fragile.
Mitigating a backlash
A common fear is that Elsevier can use its market power to create ‘lock ins’ where institutions end up paying a high fee for a low service because switching costs are too high. Lock ins are typically created through vertical integration and can go hand-in-hand with oligopolistic market structures and anticompetitive behaviour.
One should not, however, conflate market power, anticompetitive behaviour, or price gouging with assetisation. Assetisation can be one of many ways to address economic coordination problems and stabilise markets. Among economists, the tragedy of the commons is a much discussed coordination problem that refers to the challenge of distributing and protecting finite natural resources in the face of externalities. But similar challenges can arise in some situations where change of ownership is unfeasible, for example, because the thing itself is unique, very expensive to produce, only occasionally useful to consumers, or in other ways shared. Of course, it is very important to establish and discuss who gets to own assets, under which conditions, and whether the asset form is a necessary condition for the social function that the thing performs. Are assets being constructed because it helps its users, or because it helps owners?
If history is any guide, the problem arises when particular market forms or logics—in this case the logics of market dominance and assetisaiton—are pushed into the extreme and the overall market structure becomes unsustainable. Borrowing from Block’s introduction to a reprint of Polanyi’s book The Great Transformation, the instituting of particular market orders—something Polanyi discussed as ‘disembedding’ markets—is akin to stretching a rubber band:
“In this sense one might say that disembedding the market is similar to stretching a giant elastic band. Efforts to bring about greater autonomy of the market increase the tension level. With further stretching, either the band will snap—representing social disintegration—or the economy will revert to a more embedded position.” (Foreword to the Great Transformation, page xxv)
While the rubber band metaphor is meant to be applied at the level of social orders, it is a useful heuristic for thinking about Elsevier as a powerful actor in the academic publishing sector. Universities and academics alike have voiced their concerns about Elsevier by boycotting the publishers. Bad practices and its bloated size and profit margins are key reasons for discontent. The rubber band appears very stretched.
Tensions are multiplied because customers (universities and their employees) are partly paying for both the production and consumption of Elsevier’s core products (books and articles). Universities could decide to cut out the middleman and publish the work themselves, and funnel the current bloated margins into something with a better public purpose. Launched in 2015, the UCL Press provide a template for how universities can do this. Elsevier’s recent concessions in licencing contracts indicate that the publisher is aware of these larger dynamics; publishers might therefore ask themselves how they can avoid that the rubber band snaps, while at the same time protecting a lucrative market position.
Towards entanglement: balancing surveillance and service
In this video, I discuss theories and frameworks that we can use to deconstruct how edtech may try to construct and legitimate the asset form by structuring digital flows and offline behaviour. We can think about this as a balancing act. Balancing surveillance and service hinges on the Silicon Valley idea that it is legitimate to give up user generated data if we get something in return. In edtech, this typically comes in the form of analytics and personalisation, which I consider to be a concrete version of a more general move towards entangling humans and machines. The key entanglement move is between Elsevier’s digital ecosystem and its end users, which in this case materialises in the unique identifier approach.
From a technological perspective, the best way to increase the entanglement between Elsevier’s users and its digital infrastructure, would be to move completely away from PDFs and only allow users to access and work on licenced material through a reader application, such as Elsevier’s ScienceDirect PDF Reader. From a cultural perspective, this would be met with strong resistance because PDFs are ubiquitous to modern life and over time academic users have grown accustomed to downloading and collecting them. Once we download a journal article through our university library, for all intents and purposes, it feels like we ‘own’ that copy even though legally speaking that is not normally the case. We have furthermore built routines around storing, annotating, and indexing PDFs from a wide range or sources, something Elsevier knows very well through its referencing software Mendeley, which it acquired in 2013. The move away from the print journal issue and towards the PDF was fuelled by the convenience afforded by digital storage and annotation. The benefits of compulsory platform mediated access are less apparent and may appear like a new form of enclosure. In short, forcing users to give up the PDF would stretch the rubber band too much.
Morten Hansen is a Research Associate on the Universities and Unicorns project at Lancaster University, and a PhD student at the Faculty of Education, University of Cambridge. He specialises in education markets and has previously worked as a researcher at the Saïd Business School in Oxford.