Numerical sovereignty: Cytopia’s case

In this study on digital sovereignty, we focus particularly on three elements that contribute to the empowerment of the user: the world of open source, the Semantic Web, and finally the Decentralized Web. We explain how these three concepts can contribute to the emancipation of the user from the influence that their digital environment may exert over them. Beyond this introduction to these concepts, we discuss their relevance and motivate their use in software projects in general and, more particularly, in the digital format proposed for the Cytopia conferences.

A first point concerns the use of open source. This concept aims to make the source code of software accessible to everyone. Thus, anyone is able to read what the developer has actually written: for example, the server-side (or client-side) code of a website, written in Node.js, PHP, or any other web programming language, as well as the source code of an application or a web browser. Making source code openly accessible provides additional guarantees to users regarding the way in which their data are used throughout their experience across these different environments. Some closed-source programs could, for example, conceal telemetry mechanisms through which user behaviour is recorded in order to target their actions and respond to them with promotional offers. Thus, this concept allows for a high degree of verifiability of a program, both in terms of its behaviour and its vulnerabilities: if security flaws exist, a large community can propose fixes or improvements.

A second point focuses on the Semantic Web. One way of revolutionising our interaction with the Web would be to make it more semantically accessible. The current Web essentially consists of websites whose content is formatted according to the specific purposes and interfaces of each individual website. Within the concept of the Semantic Web, the goal would be to consider the Web as a more unified environment, in the form of a giant database. Several approaches allow us to transform our current vision of the Web into a Semantic Web. For example, with the advent of LLMs (Large Language Models), we can now browse a website through an LLM that serves as an intermediary, interpreting its content and presenting it according to the user’s perspective. This method nevertheless remains costly, as it requires significant intervention from these language models. Other methods make web content more accessible to these LLMs, such as semantic indexing through vector databases or the use of semantic graphs.

Thus, the Semantic Web constitutes an additional approach towards the emancipation of the user from a heterogeneous digital environment, where navigating through different websites requires an additional cognitive effort to adapt to each new environment. The vision of a Web equipped with a single conversational interface providing access to different sources of information would allow for a better understanding of information and greater accessibility for the user.

A final point focuses on the Decentralized Web. This concept focuses on the network architecture underlying the Web. The dominant model on the current Web is based on the centralisation of data: users navigate through dynamic websites where their data are collected, processed, and stored on a server. This confinement of data also imprisons the user: they are obliged to continue using the services associated with the servers that have collected their data and may potentially be encouraged to pay a subscription in order to store even more data.

The principle of the Decentralized Web aims to eliminate this lock-in effect: servers would no longer be responsible for storing data, but would instead serve as channels for transmitting information without necessarily being able to read it. Thus, the only holders of the information would be the users themselves.

Following these three points, we apply these principles to a specific use case: the Cytopia.fr website. In its final form, the latter will attempt to integrate these three principles as effectively as possible in order to provide greater sovereignty to the user. An open-source format will be adopted in order to make the project’s source code available for consultation and replication. The Semantic Web will also be implemented in order to make the website navigable through an LLM, particularly an LLM that runs locally on the user’s device and is open source: we thus avoid the use of proprietary and paid LLMs.

Finally, another, more distant objective of Cytopia would be to make its knowledge accessible to users by allowing them to capture it, structure it, and store it locally on their own devices. Thus, information would not belong exclusively to a website: the latter would be able to distribute it without requiring users to remain dependent on the original website.