A Retrospective Look at the Future of the Data Economy

Once my doctoral thesis has been finalised, it is worth reflecting on where the data economy is heading to and what future works may be worth developing based on the findings and conclusions of my PhD. During the defense, I was asked several specific questions about this, and I am leaving here some thoughts on relevant topics that we discussed after my dissertation.

First, I learnt that we loosely talk about «data» and «data sharing» when there are many categories of data useful for a number of use cases with very different requirements. Data pricing, packaging, quality metrics or delivery methods may differ substantially depending on which kind of data we are dealing with and for what purpose. Think of, for example, high-frequency trading data compared to historical weather data to train forecasting models. This is why I think that specialization is key to winning in the data economy.

On the contrary, most pioneering data marketplaces intended to generalise and aimed to trade any kind of data. In time, more specialised platforms have appeared which are able to provide more added value because they focus on a specific category, service or use case. Furthermore, in our survey paper we talk about the «commodification of data exchanges», meaning data exchanges and private marketplaces will likely become a functionality of data management platforms and digital services, which will coexist with standalone marketplace platforms.

Even if defining a general data exchange workflow is not straightforward and may need to consider the specificities of some use cases, I think we must seek for standard data exchange protocol. What is not clear to me is whether it should be a de jure or a de facto standard.

In this direction, some initiatives like Gaia-X and International Data Spaces are looking to set up a common governance framework, and some start-ups (Ocean Protocol) and data management systems are also defining their own workflows. Not only will this help address the existing market fragmentation – providers in need to deal with a number of platforms -, but it will also will help in controlling and measuring the data economy. And doing so will be key as the global economy becomes increasingly digital and data-driven.

The motto of the PhD was an impactful book by Jaron Lanier that advocates for paying people for their (data) contribution to the digital economy. Consequently, I spent some time studying Personal Information Management Systems (PIMS) and I reflected during the PhD about what would be the case for such entities in this context. My conclusion is that there are opposite and contradictory trends affecting the business of PIMS.

On the one hand, we have observed a a growing concern about privacy and the impact of AI in the economy and the labor market in the last four years. What about workers training models that may ultimately render them unnecessary in the market? Maybe at the beginning of the PhD this idea of getting paid for the data we produce seemed crazy, but in the last four years it is gaining momentum and news about developments and the impact of AI on the society are in mainstream newspapers every week. This context may pave the way for PIMS to assume a pivotal role in managing the information of individuals in such a human-centric data economy.

On the other hand, I do not think there is a significant willingness to pay for their services. Moreover, business models of pioneering PIMS have not shown to be standalone viable. Half of the PIMS in our survey ceased operations or significantly reduced or shifted their efforts during the course of the PhD (e.g., Wibson, DataCoup, etc.).

In my opinion, PIMS are meant to become a sort of (data) unions integrated with service providers. We often compare data to other production factors like capital, land, or labor. What do these key economic factors have in common? They are all heavily regulated markets, and so will be data markets in the future. Amongst data regulations, maybe (sufficiently big) service providers might be obliged to integrate PIMS or funcionality alike to protect the rights of individuals using their services. Such unions can also expand to also supervise the compliance of AI regulations and detect biases and bad practices just as labor unions and consumer protection organizations do in the traditional economy.

Even though the PhD was not especially devoted to studying existing regulations or developing new ones, I was also asked during the defense what key regulations would be needed in this context. A tough question when there is already a siginificant ongoing regulatory effort in the EU including the General Data Protection Regulation, the Regulation for the Free Flow of non-Personal Data and guidelines, the Data Governance Act and the Data Act. There are other policy initiatives to create a common data space in the EU Building a European Data Economy and the European Strategy for data as key pillars of the strategy for “shaping Europe’s digital future«. Related to these, the policy about “Artificial Intelligence” with the strategy for “Artificial Intelligence for Europe”, and policies towards ensuring EU autonomy in supplying European cloud services.

My gut feeling is that governments and regulatory bodies are often half-blind when making policy and regulatory decisions on these topics, Consequently, I would say that initiatives are needed first to increase the transparency and set up mechanisms to better measure and understand the new data economy. Some artifacts of my PhD were aimed to that direction. Moreover, this is becoming urgent, as ongoing initiatives like the ‘data dividend’ in California, the ‘data tax’ in NYC, or digital service taxes (DST) in Europe will soon require to put a financial value on data.

Bottom line, this journey started almost 4 years ago, when I came across a crazy PhD offered by Prof. Nikolaos Laoutaris at IMDEA Networks that wanted to pay people for their data as a solution to the privacy problems on the Internet. Having witnessed the evolution of AI and the growing importance of data in this period, maybe that was not such a crazy idea. On the contrary, news and facts have remarked the importance and urgency of more research efforts on the value of data and of more multidisciplinary PhDs that address this relevant topic.

Thanks for reading!