How a united Europe could democratise data
A short essay on a hypothetical future of personal data
Introduction
Due to recent events, the topic of a united Europe has gained new momentum.
I am of course referring to the conflict in Ukraine and how it moved the dependence of Europe on autocratic regimes under the spotlight. But there were also supply chain complications during the Covid-19 pandemic which have revealed a similar reliance on Chinese manufacturers.
It is interesting to see how this has resulted in European countries appearing more united than they have been for a long time. This gave me hope that they could work more closely together in the future and I wondered what could be achieved together. Could we use this momentum to combat other challenges of our time, one of them being big-data companies threatening democracy?
Because in addition to the events noted earlier, we are giving up another form of independence by surrendering our data into the hands of private companies.
In this day and age, especially with advancements in machine learning, whoever controls the data controls the people. The recent years have revealed how artificial intelligence, particularly on social media, can have a significant impact on how our society — and therefore democracy itself evolves.
I believe that Europe, due to its history, location, and economic power is responsible for maintaining a robust global democracy.
Therefore, I want to discuss a new approach to dealing with the personal data which every one of us produces, but it requires European countries to work together.
Technology
The approach consists of the combination of two concepts, the first one being federated learning and the second one being data trust centres [1].
Federated learning, as one form of machine learning, has been around for a long time and already found several applications.
In a nutshell, instead of the data being brought to the machine learning algorithm, the algorithm “visits” the data on an external server.
Why is this useful? Well, it would allow us to keep the data of European citizens in Europe, while still providing tech companies with the data they need. Therefore, European regulations can be utilised to enforce privacy policies et cetera.
But how do we store all this data?
This leads us to the second concept: decentralised data trust centres.
Currently upon visiting a Website, we often contractually permit its owners to store our data. With the approach highlighted in this post, the user would instead have a contract with the publicly owned trust centre. It could contain which companies are allowed to access and store the data and even what kind of data can be stored, essentially giving the user the choice over their data.
The lack of this choice is one of the main issues in our data-driven world.
Aside from these personal benefits, there are also significant socio-economic benefits that come with decentralised data trust centres.
Firstly, it would enable a more diverse application of artificial intelligence across industries. Often it is the data which holds the power, not the machine learning algorithm itself. Making the data available to other companies provides small and medium-sized businesses with the opportunity to improve productivity by utilising artificial intelligence themselves and thereby creating new jobs.
Going hand-in-hand with this is the diversification of interests backed by big data. As of right now, the few companies which have access to large pools of data often act with specific corporate and ideological interests in mind. The hope is that this will change once the data is made more accessible. Additionally, it could significantly improve healthcare research, which suffers from a severe lack of medical data, especially of women and children.
Lastly, decentralised means that the data is not stored in a single place, but spread out on servers across Europe. No single country would have control of the data, and if one server fails the system would still be functioning.
Final thoughts
To put it briefly, we could achieve a balanced application of AI by democratising data. This could help regulate monopolies and strengthen the diversity of data-driven applications.
All of this would be beneficial for democracy and nurture the independence of the European Union.
Still, many people are afraid of these large-scale changes and it often lacks the political constraint to put such plans into action, which is why I believe that only a united Europe would have the tenacity to do so.
____
Thanks for reading! Hopefully, this post provoked some interesting thoughts.
I probably didn’t explain everything perfectly, but I’ve been thinking about the concepts mentioned above for quite a while and wanted to write something about it.
Resources:
[1] A (German) Podcast by Handelsblatt Disrupt with Katharina Zweig deals with data trust centres and federated learning.
It was my main source of inspiration for writing this article.
