Among many objectives, the DGA aims to provide a European alternative to the current market-based approach dominated by a few handfuls of actors – we previously commented on the final version of the proposal. As emphasized in recital 25, the measure aims to provide a “novel European way of Data Governance” by devising a new data sharing framework based on ‘data intermediation services’ to allow individuals and companies to engage in data sharing in a secure environment.
These data intermediation services must be structurally separated from any ancillary commercial activity and cannot use the shared data to improve their own intermediation services. The Commission hopes that by introducing such a new market structure for data sharing services, Europe’s dependence on Big Tech’s platformized services will decrease through the proliferation of independent data intermediaries in the digital single market.
However, a closer look at this new market structure reveals an inherent contradiction; namely, that a novel European way of data governance still hinges on a platformized business model for data sharing. This is because such an intermediation model relies on a central intermediary acting as a centerpiece on a multi-sided platform. In this case, between data subjects and data holders, on the one hand, and data reusers, on the other hand. And such intermediaries, as it became clear in the last decade, have a tendency to centralize their power.
In this light, data intermediation services do not seem to substantially alter the dynamics of informational capitalism, but rather simply provide an alternative to current market practice with a platformized business model. This confirms Julie Cohen’s – Professor of Law and Technology at Georgetown Law – claim about the inevitable platformization of data markets: “everything old is new again”.
In the DGA, this tendency towards platformization might be the result of two regulatory choices: those related to the commodification of information assets, and the potential recentralization of market power in data intermediaries’ hands.
Data intermediation services are a service, which aims to establish commercial relationships for the purpose of data sharing between an undetermined number of data subjects and data holders, on the one hand, and data users on the other hand, through technical, legal or other means, including for the exercise of data subjects’ rights in relation to personal data.
As it clearly emerges from the definition in article 2, data intermediation services establish a commercial relationship between different user groups. This broad definition opens the door to a variety of approaches related to data intermediation, including personal information management tools, such as personal data spaces or data wallets, but also broader mechanisms to make data available for free to the general public. The Commission left this definition open-ended in order to allow wide experimentation in this field.
Yet, the baseline approach does not change across various potential use-cases: being services of commercial nature, data intermediation services will allow for data sharing situations where information is exchanged as part of one user’s property. This confirms a vision based on a propertarian approach to data governance where data is mainly seen as someone’s by-product of online activity, to be exchanged in return for a service or compensation. Such a conceptualization however omits the relational component embedded in data production as well as its non-rivalrous characteristics, which are opposed to current market practice for information ownership.
So-called “services of data cooperatives”, classified in the DGA as a sub-category of data intermediaries, are a potential exception. In theory, data cooperatives provide a framework for the collective exercise of users’ data rights rooted in collective ownership of data and shared decision-making among members. However, the text reads vaguely on this possibility as data cooperatives would mainly have the “principal object to support its members in the exercise of their rights”.
A recent paper by Elettra Bietti, Ander Exteberria, Morhsed Mannan and Janis Wong – published in the context of Harvard’s research sprint on cooperative data governance – tries to provide more clarity on this issue. According to the authors, data cooperatives, in current DGA phrasing, “do not challenge the economic unfairness of data capitalism” as recital 24 highlights that GDPR rights are personal rights and therefore cannot be waived by data subjects. This seems to suggest a reading where no real delegation or conferral of rights to a third party can occur for users’ personal data as processing always requires the exercise of users’ consent at the individual level.
However, such interpretation would necessarily imply that collective data governance for personal data is not possible under GDPR, even if it ultimately empowers consumers. This would however be in sharp contrast to European data protection goals which aim to enhance user self-determination and autonomy in the online sphere. More clarity on the mandability of data rights is needed.
In addition, such a strong reliance on the data intermediaries’ model might give rise to scenarios of market concentration that the European strategy for data is supposed to be fighting in the first place. After all, data intermediation services are inherently multi-sided platforms that are likely going to enjoy significant network effects across data users, data reusers, and data holders. Of course, they cannot use the intermediated data to improve their services given strict regulatory requirements to ensure their independence from any for-profit activity.
Yet, such an open-ended definition opens venues for data monetization as intermediaries will be able to execute transactions across user groups. And it is in this context that the question of incentives arises, especially for those data intermediaries that will provide sharing services free of remuneration in the common interest: will they be able to sustain their business model without resorting to some kind of data monetization?
Although it is relatively too early to properly assess the feasibility of remuneration-free schemes, it is evident that intermediaries which would simply make data available in the general interest would significantly struggle vis-à-vis for-profit services. In this light, the DGA introduced the idea of data altruism organizations to make sure that users could voluntarily and freely donate their data for the common good. However, as criticized by Winfried Weil in this discussion paper commissioned by Algorithm Watch, data altruism seems to be “a missed opportunity to breathe life into the ‘data for good’” idea. This is largely due to the burdensome requirements related to setting up a data altruism organization as well as for restrictive GDPR data processing obligations that hinder the processing of personal data for altruistic purposes.
The DGA will provide a new alternative to counteract big tech dominance in data markets. Nevertheless, current provisions regulating the rollout of data intermediation services in the digital single market seem to favor platformized services that will act on a for-profit basis. This could reinforce a platformized market scenario where much of business power would lie in the hands of a few intermediation services instead of fostering decentralization.
An alternative could be provided by kickstarting a space for non-commercial intermediaries that would act in the public interest for purposes of making data available as a commons. For instance, Nick Srnicek – lecturer in digital economy at King’s College London and author of Platform Capitalism – advocates for the creation of public platforms which could redistribute data through collectivism. In this light, the Digital Europe Program (DIGITAL) could provide much-needed public funding to sustain these alternative business models, as part of its underlying goal to create common and interoperable data spaces. This step is necessary if we want to create a novel European way of data governance instead of relying on existing, platformized business models.