2025 seems far away. But 2025 isn’t a science fiction perspective, with robots taking over the world. 2025 will arrive sooner than we think with massive impacting digital changes. Especially when it comes to data science. By 2025, all our digital capacities will have doubled twice. The world might not be ready for quantum computing just yet, but the sources of data will be vastly larger.
The Graduate Institute of International and Development studies (IHEID) and the EPFL Center for Digital Trust (C4DT) have decided to organise the Data 2025 conference to start the conversation and help create a community to reflect on issues related to data.
The definition of data, healthcare data, and data governance are some of the aspects that will be covered by the panels. To pave the way to a better understanding, IHEID Professor, Richard Baldwin, and C4DT Academic Director and EPFL Professor, Jean-Pierre Hubaux, have agreed to share their experience as a preview.
What is data. Data can only be defined in a context, Professor Baldwin explains. Data is information that applies to one question that is organised and recorded for many instances. In other words, what data is in one context, might just be noise in another.
“Data is a group of information that is useful for particular purposes.”
From the general definition of data, one wanders to the more specific question of personal data. The notion is connected to time and exclusivity dimensions, Professor Baldwin continues. Personal data is related to aspects of one’s life at a certain moment of time and is not available to the general public.
Personal does not necessarily mean sensitive, however. Professor Hubaux relates sensitive data to the amount of damage that can result from its misuse. In the case of healthcare, for example, it means a loss of trust between the caregiver and the patient, which is at the core of an efficient healing process. This makes protection of data particularly important as it refers to a historical tradition.
“Just go back to the Hippocrates oath, more than 2200 years ago. One of the statements is about confidentiality of the data. It is at the core of the trust, which means that on one hand, there is a willingness to make use of medical data of as many patients as possible to improve medicine. But on the other hand, we cannot just open all the files, it would be unacceptable. We have to live with this tension.”
A tension that revolves around a sacred and highly discussed notion: privacy. But not only. To privacy, one must add economic value and property concerns. And Covid has exacerbated the debate.
Data and health. The pandemic certainly revealed that the world was unprepared on many fronts, in particular because it does not have the proper data pipelines to make quick and meaningful decisions, Professor Hubaux continues, especially when it concerns health.
“Where are the electronic health records? We don't have such a thing. I think we need to move way more aggressively on that front. Everyone has understood the relevance of an efficient healthcare system and the ability to rely on quality data.”
And when it comes to health data, the situation is more complicated than in any other field. Medical data is of a multiple nature. It can be a diagnosis, instrument measurements, a patient’s self-reported issues, clinical data, genetic data, the examples are numerous. But accessing health data can contribute tremendously to move research forward. The question is how.
“What is missing nowadays are the incentives for data sharing. We need to overcome these barriers. And a project such as MedCo does exactly that. It has the ability to make computations without transferring the data, and solving the problem of the protection as a consequence.”
An operational system that makes decentralized medical data available for research and healthcare with the highest protection standards, MedCo is a successful example on how data can be shared in a secure way. it has been developed by the EPFL Laboratory for Data Security, in close partnership with the Lausanne University Hospital. MedCo contributes to fully understand the pathologies and define adapted treatments.
A hopeful result that could perhaps inspire actors on a global level, to reach an optimal data governance system.
Data and governance. The Data 2025 conference is precisely about starting a conversation from a broad and multi-stakeholder perspective, as part of a global digital governance process. Professor Baldwin explains:
“To have a robust discussion on what data is, what it means and how it could be governed, is very important. It is important to get a better understanding of the multidimensionality of data, and how this digital technology has changed what data means and its role, not just in the economy, but in our private life, and even political life. Data is becoming part of statecraft. Before we felt that the conversation was too fragmented and too siloed. And the people didn't understand each other's perspective. People in governments tend to not understand the technical aspects of it enough. And the companies tend to ignore the governance aspects of it.”
If data can only be fully understood in context, it certainly seems indispensable to share this context with others in order to reach new levels of cooperation and look forward to the opportunities today already. Because 2025 is just around the corner.