Your data is worth nothing
We need a more nuanced understanding of how personal data is transformed into a valuable asset
Personal data is a key ingredient in the current economy. By now everyone is aware that “free” internet services are not free at all, and that the data generated by browsing online is processed and monetized by tech companies in a rush to improve content targeting and refine techniques to foster our consumption of things that, in general, we do not really need.
So, the cat is out of the bag, and yet, it is hard to cut through the fog of a current debate characterized by a combination of confusion and cliché. Making sense of this noise is essential if we want to design responses to the harms created by data without foregoing the benefits that it also carries with it.
One of the things that gets confused is the concept of data. For example, personal data is just one type among the vast amounts of data generated by new technologies. Weather data (or traffic data) is not personal data and can still be collected, processed, and used to generate value, but the challenges that it poses (i.e., if this data is made available by public institutions, are the companies building products and services on top of it compensating public institutions for some of the investment that went into its creation?) are different from those coming from personal data. Understanding what data we are talking about is essential to identify their specific challenges and associated responses.
Most of the debates around the challenges of datafication do have – whether they make it explicit or not – personal data in mind. Researchers, journalists, and policymakers are broadly aware of the uses and misuses of data that can result in surveillance, political meddling, violations to privacy, entrenchments of biases, discrimination, and economic inequality. Given the diversity of these problems, however, it would be naïve to think that a single solution could address all of them, and yet often one can read someone list all these problems to immediately propose a silver bullet that will, purportedly, solve them all at once.
One such solution is that of creating individual property rights over personal data. This, the theory goes, will give each individual more control to decide who to share the data with and in what terms (for example restricting access to some data by certain companies or asking for an economic compensation for such data). This idea of generating property rights was first advanced by the World Economic Forum a while ago in the report “Personal Data: the Emergence of a New Asset Class.”
Legal scholar Salome Viljoen recently developed an interesting critique to this approach to data governance (here the academic piece published in the Yale Law Journal). She argues that the individualistic approach to data governance ignores its relational dimension, which is what drives much of the social value and harm of data. Since data collectors are highly concentrated and data subjects fragmented and isolated, it is not clear how granting individual property rights will tackle that power imbalance. Unless there is some sort of collective bargaining mechanism that groups together data subjects (now data owners), it will be difficult for an individual to exert much influence or even exercise their rights. Viljoen also makes clear that property rights, a measure mostly conceived to tackle economic concentration in the exploitation of personal data, will also hardly address concerns regarding privacy or other political and social rights:
Paying data subjects at the point of collection does nothing to address uses of such data that may violate the civil rights of others and amplify existing social oppression. In fact, by legitimating the marketplace for data, payment may legitimate downstream practices that result. Payment may also incentivize people to share data about themselves and thus further degrade privacy.
Another important criticism to the proprietary approach to data comes from sociologists and web researchers Thomas Beauvisage and Kevin Mallet. In Datassets: Asseting and Marketing Personal Data, they argue that the ability to generate value out of personal data is the result of a process of assetization, which, as described by Kevin Birch and Fabian Muniesa, refers to the technical and legal mechanisms to transform something (e.g., personal data) into an asset that can be “controlled, traded, and capitalized as a revenue stream.” Many companies have tried, and failed so far, to create consumer-to-business markets for personal data. A key reason behind this failure is that “on existing markets, individual level information is not worth much, and this effective market price probably does not justify individuals going to the effort of putting them up for sale”. They go on to conclude that “personal data cannot be considered as good as-is but need to be constituted as products to match operational needs and take a specific place in market architectures.”
The paper describes the market architecture of business-to-business exchanges around the needs of advertisers. This market pre-existed the emergence of digital platforms, but has been turbocharged by the dominance of these platforms and the widespread use of automatized algorithms. In this market personal data is not directly sold and purchased as a commodity, but rather, it is transformed and repurposed in an assetization process driven by intermediaries and market makers. Without the products derived - but distinct - from personal data that these market and entrepreneurial actions generate, individual personal data is worthless.
The implications of this are clear: governing (personal) data requires understanding how this process of assetization works to clearly identify where in the process are rights violated and harms generated. More broadly, the debates about data governance demand more clarity in the use of terms, identification of problems, and design of possible solutions.
A few simple questions may help in providing that clarity. First, what type data are we talking about? Second, what is the problem that we are trying to tackle? If the problem is about market concentration and value extraction of personal data, a third question is: how is personal data transformed into a valuable asset in each industry? Fourth, what stages in the data collection, processing (including assetization processes) and use can be influenced or shaped to reduce risks and harms? Fifth, what are the models or policies that can do so while maintaining the benefits of using data?
True, this is a relatively new topic, but it is time for the public debate to become more nuanced and accurate when it comes to discussing the promises and perils of our datafied society.