Facebook has been in the news a lot recently, and rarely in a good way. This post is not about Facebook, but it is about that business model, critiqued brilliantly by Jaron Lanier in a Politico post last week. It is clear now that it is possible to have very negative effects with a business model that introduces perverse or anti-social incentives into a global communications system. The essence of the business model is that users are not the customers; the customers are those exploiting the users’ data.
Are such business models in place in the academic research world? Well, yes probably. The main academic social networks, Academia.edu and ResearchGate are free for users, and have attracted commercial investment on the basis that they will make money in other ways. Those other ways may now be under scrutiny as never before. What would be the academic research equivalent of Cambridge Analytica’s alleged activities based on Facebook data? Perhaps commercial R&D companies would buy insight into research trends or groups, which might give them a commercial advantage and improve outcomes for the economy and society – a great thing, providing the users of academic social networks knew the full extent of the data they were giving away. Or perhaps that insight could be used to lobby governments, or research funders, or university leaders, or undermine research unfavourable to a particular product or industry or political party?
And, shouldn’t information about academic research be available first and openly to the research community? That seems, at least to me, to be an inescapable implication of the Mertonian norms of science.
But is the problem wider? The reason that the free-to-use social network business model works is that the network is a walled garden. It is a central service in which virtually all interactions are internal to itself, exactly the opposite of the original vision of the internet and the web. That is an extreme form of lock-in, but there are other kinds, helpfully reviewed earlier this year by Roger Schonfeld. The recent announcement that Clarivate has acquired Kopernio only draws more attention to the three research workflow environments that are emerging as the main players in a new market – Clarivate, Elsevier and Digital Science. At present these environments are not as integrated as the features of Facebook or ResearchGate. Therefore, they may not yet be able to offer researchers the kind of experience that enables the platforms to collect the kinds of user data that would underpin a free-to-use business model.
And yet this presents a quandary to those of us committed to open research, where by “open” we mean, among other things, free to use. What business models are unlikely to present notable risks to open academic research, if the research itself needs to be open and selling user data is fraught with difficulty?
All of this has alarmed some researchers sufficiently to set up a community-owned alternative, the ScholarlyHub. There is also a rapidly growing group of community-run preprint services and research funder publishing platforms running open peer review and other social functions. And reports such as the Metric Tide call for an open data infrastructure to allow information about research to be shared within the community. This suggests that at least some senior and influential members of the academic community see a problem to be urgently addressed.
Last week I attended a workshop at the Wellcome Trust on meta-research, which of course needs evidence on which to work. It would seem odd if the research community had to divert money away from research in order to pay for data about its own activities; and yet how else to fund the curation of those data? Perhaps some of this comes back to the collective action problem I referred to in my last post.