After years of discussion and planning, the Digital Services Act was approved by the European Parliament in July 2022. It has many provisions for intermediary digital services (social media platforms and online marketplaces). Some of these provisions remain hotly debated. But everyone can pretty much agree on one new obligation: transparency and data access requirements. One key element of those will give researchers access to key platforms’ data in order to scrutinize how they work and better understand how online risk evolves.
Indeed, four years ago, the transparency reporting requirements were the only details in Germany’s Network Enforcement Act (Netzwerkdurchsetzungsgesetz) that everyone, from international rights organization ARTICLE 19 to the law’s most ardent supporters, could get behind — as Paddy Leerssen and I showed in an analysis of that law’s attempt to regulate content on large social media platforms. And the necessity of transparency remains a key point of consensus for other countries planning regulation of digital services, including Canada.
The idea is, obviously, that when researchers can access more data about, say, algorithmic processes, their research can generate findings that will provide the basis for evidence-based policy. Everyone can agree on the importance of access; the main dispute is about whether the definition of a researcher should encompass those not employed at universities, such as people working at digital rights organizations or think tanks. (One example of grappling concretely and deeply with enabling research access is the report from the European Digital Media Observatory’s Working Group on Platform-to-Researcher Data Access.)
A less discussed but equally pressing issue is that for all the praise of transparency there has been no conversation about a real and central problem in the research community of higher education: as universities become ever more reliant on adjunct teaching, will there be enough researchers to actually conduct this work? The crisis in higher education could undermine the drive toward transparency and accountability that governments seek.
For those outside the academy (and especially those inside it), the statistics paint a grim picture. The vast majority of Ph.D. students will not find employment as tenure-track professors. In North America, only around 10 to 20 percent of Ph.D. students become professors. This holds for biochemistry graduates from University of Toronto, English graduates from McGill and history Ph.D. graduates in 2019–2020 in the United States. Many others remain in higher education, but as administrators or adjunct professors working on short-term contracts. Their employment would not provide much time or space for evaluating mounds of data from Meta.
Some might argue that these statistics merely reflect that more Ph.D. degrees are being awarded. But the number of professors does not bear this out. In fact, the overall professoriate will start to decline quite rapidly as people retire. One of the main messages of a 2021 report from the Council of Canadian Academies is that “the number of tenure-stream professors in Canada has been almost constant since 2009, with a declining number of assistant professor positions — normally the entry position for new faculty. In 2009, there were more than 10,500 assistant professors in Canada, but by 2017 this had fallen to about 8,600 as universities did not fully replenish their ranks after promoting assistant professors to associate professor positions.”
The number of Ph.D. graduates may be increasing, and they can certainly bring invaluable labour to bear during their studies. But without long-term faculty who have the time to build research profiles, labs and agendas, all the transparency in the world will not guarantee better outcomes.
I have also deliberately drawn statistics from multiple disciplines because any work on platforms has necessarily become interdisciplinary. While computer scientists and some communications scholars engage in quantitative work, many insights from transparency data could emerge from other disciplines such as English, law or history.
Back in March 2021, Daphne Keller of Stanford’s Cyber Policy Center suggested that we needed more “humility about transparency.” She noted that “researchers and civil society should assume we are operating with a limited transparency ‘budget,’ which we must spend wisely — asking for the information we can best put to use, and factoring in the cost. We need better understanding of both research needs and platform capabilities to do this cost-benefit analysis well.” Even beyond research needs, it is time to recognize research limitations in our current higher education landscape and think about how to address that problem too, if policy makers expect researchers to use newly available data to understand platforms.
Ironically, it is quite probable that platforms have hired many of those excellent doctoral students who could not find jobs in academia. Already in 2017, 16 percent of jobs at Google required a doctorate. If the companies have far more researchers than universities have, transparency regulations may not do as much to address the imbalance of knowledge as many expect.
It won’t work to place the burden for evidence on the research community if poor funding continues to erode the foundations of that community. It may sound like special pleading for higher education for a professor to point out the problems with universities that are not hiring enough professors. But the problem needs to be aired before politicians and policy makers place expectations on researchers that researchers cannot fulfill.