When the readers of The Scholarly Kitchen last heard from me on the about ResearchGate, I suggested that ResearchGate might emerge unscathed, perhaps even strengthened, from attempts by publishers to tame it through take-down notices and lawsuits. Though the take-down notices continue and the litigation is still ongoing, the recent announcement of a negotiated agreement between ResearchGate and Springer Nature, as well as Cambridge University Press and Thieme, indicates that ResearchGate may be proving its staying power in this field.
Last October, talks between the the International Association of Scientific, Technical and Medical Publishers (STM) and ResearchGate dissolved. Subsequently,the Coalition for Responsible Sharing (CRS) was created, which has pursued a strategy of take-down notices as well as continued discussions with ResearchGate. CRS includes Elsevier and the American Chemical Society (ACS), which are pursuing litigation against ResearchGate. Shortly thereafter, a very succinct press release appeared. In totality, it said:
ResearchGate and Springer Nature have been in serious discussions for some time about finding solutions to sharing scientific journal articles online, while at the same time protecting intellectual property rights. The companies are cautiously optimistic that a solution can be found, and we invite other publishers and societies to join the talks.
I pondered at the time about what exactly this might mean. I was particularly curious about what concessions Springer Nature and/or ResearchGate would make relative to the original STM demands, those of the CRS, and the claims in the Elsevier/ACS lawsuit. With the recent press release announcing a collaborative agreement between ResearchGate and Springer Nature, Cambridge University Press, and Thieme (collectively “SNCUPT” for the remainder of this essay), we can now start to see the divergences and speculate on the implications of those differences.
I am grateful to the staff of ResearchGate, Springer Nature, Cambridge University Press, Elsevier, and CRS for taking time to talk with me and share their perspectives on the agreement.
The Primary Issue — Pre/Post Upload Screening
The fundamental difference between SNCUPT and CRS is willingness/unwillingness to accept what seems to be ResearchGate’s line-in-the-sand principle — no automated copyright compliance check at the point of user upload. It almost seems like ResearchGate might be asserting that it is not feasible to do this review at upload because author agreements are not available for ResearchGate’s review. As Ijad Madisch, CEO of ResearchGate, explained to me in an April 27 email interview, “when scientists publish in a journal, they sign complex, private licensing agreements with the publishers. We’re not privy to these agreements that determine whether scientists can upload an article to ResearchGate and under which conditions. Some allow it, others don’t.” Given ResearchGate’s actions to turn content private in light of the planned CRS take-down notices last fall, one may have to question this assertion; however, it appears that ResearchGate has prevailed in its position in the SNCUPT agreement.
So, under the SNCUPT agreement, content will be made available on ResearchGate without pre-screening. However, also under the SNCUPT agreement, ResearchGate will facilitate publisher ability to review the content that has been posted and pursue various options for reacting to infringing content. Publishers of course retain the option of issuing take-down notices and ResearchGate will comply (as they are legally required to do so regardless of the SNCUPT agreement) but the agreement also includes a commitment to developing other pathways. Options here might include making the content private on ResearchGate, which still allows the author to share it upon request, replacing an uploaded PDF with a inline reading copy (e.g., perhaps through Digital Science’s ReadCube or Cambridge Core Share), or engaging authors in a process of education through which the authors are given options among which to choose. It is particularly notable that linking out to a copy on another server is a possible option as ResearchGate has not allowed individual users to do such linking out under its terms of service.
James Milne, spokesperson for the CRS, reiterated that this approach of post-upload screening is unacceptable to the CRS. Milne again emphasized that the group looks for a simple solution that automatically provides clarity and allows for a consistent user experience without disruption. Milne outlines that the CRS believes that ResearchGate has an obligation to deploy copyright infringement detection software proactively. Likewise, this is the very issue that Elsevier/ACS seek to have the courts address. CRS continues to offer ResearchGate software that would do such proactive screening at upload in a negotiated agreement that includes pre-screening.
Gaby Appleton, Managing Director of Mendeley, detailed for me how the copyright infringement detection screening works. In a nutshell, a PDF is ingested, features (metadata, etc.) are extracted from the PDF, and then processed through a decision engine. The goal is high precision and low false positives. This Elsevier technology is part of the process through which CRS has been identifying infringing content in order to send take-down notices to ResearchGate.
The CRS also insists that the private sharing feature of ResearchGate is not an acceptable approach for authors sharing with their work because it is not constrained to a preset research group as the STM Voluntary Principles on Article Sharing on Scholarly Collaboration Networks allow, but rather functions as a mechanism to share an article with anyone who might ask for it. The SNCUPT does not address this specifically but allows each publisher to define the choice they might pursue if content is uploaded that infringes on publisher copyright ownership.
The Value of User-Centered Sharing
The STM Voluntary Principles are grounded in the fundamental principle that article sharing is important to the advancement of research:
Scholarly research is by its nature collaborative. Teams of researchers and scientists in the academic and not-for-profit sectors share experience, expertise, and facilities in order to advance human knowledge and understanding. Nowhere is this more evident than in the sharing of scholarly articles.
Brigitte Shull, Senior Vice President of Academic Publishing, Americas and Director of Scholarly Communication of Cambridge University Press, said it even more simply to me when we spoke: “sharing is part of the research workflow.” Shull explained that Cambridge understands that their authors value ResearchGate, not only for dissemination of and access to content, but also for collaboration. ResearchGate is not only a space for sharing articles but also for creating connections and conversations.
Listening to Shull describe Cambridge’s work to understand author motivations and drivers made it clear that taking a user-centered approach to the “problem” of ResearchGate rather than a copyright ownership-centric approach brings a different lens to the decision-making around negotiations with ResearchGate. Copyright issues are still addressed in the SNCUPT agreement but in a way that centers researcher workflow.
Steven Inchcoombe, Chief Publishing Officer of Springer Nature, echoed similar themes in stating that “we aim to make the content we publish as discoverable, accessible, understandable, usable, reusable, and shareable as possible and a part of that means putting our content where researchers already are rather than getting them to come to our content.” As we chatted, Steven went on to say that “ResearchGate is an important and popular place where this can all occur” and “we want to partner with anyone who serves researchers — through our respective strengths.”
In addition, through the SNCUPT agreement, publishers will have access to usage data about how their content is engaged on the ResearchGate platform. These data can be integrated with other usage data to provide authors with total usage metrics about their work and publishers will be able to compare/contrast content performance on the publisher platform with content performance on ResearchGate. This is a point of convergence with CRS, the members of which would also like to have usage data about content on ResearchGate, but are unwilling to trade-off the demand for pre-screening of content as the SNCUPT agreement requires. Additionally, the Elsevier/ACS lawsuit, even if successful, will not compel ResearchGate to share usage data. If SNCUPT publishers are able link out from ResearchGate to copies hosted on their own platforms, they will also be able to capture usage data of their content directly.
This is not the final chapter in the story of the relationships between ResearchGate and various publishers but this negotiated agreement with SNCUPT does demonstrate that there is not a uniformity of perspective in the publishing community about article sharing on ResearchGate, or presumably on the many other scholarly collaboration networks that exist. It also signals that ResearchGate, a decade-old start-up disruptor with with venture capital investment and a rapidly grown user base, has taken its place at the negotiating table and found not just enemies but allies.
The SNCUPT agreement lends a kind of legitimacy that begins to distance ResearchGate from piracy sites like Sci-Hub. Presumably discovery services will need to reconsider any decisions that marginalized ResearchGate hosted content and those who do research on the extent of open access content will need to grapple with how to categorize ResearchGate content (will it be considered “bronze” when there is a ResearchGate-publisher agreement in place or will yet another color be added to the lexicon?). And, the question that has been asked in conference hallways and receptions has re-emerged. Will ResearchGate remain independent or will it formalize its alliances through a merger or acquisition?