Modeling deceptive design mechanisms in digital services

Presentation

Flora Brochier began her thesis in June 2025 on the subject of modelling deceptive design mechanisms in digital services.

Her thesis is directed by Bruno Bachimont (Costech), Serge Bouchardon (Costech) and supervised by Karl Pineau and Frédérique Krupa (L'École de design nantes atlantique).

The thesis project was funded by the Fondation de France.

Why this thesis project?

Deceptive design mechanisms in digital services, often referred to as "dark patterns", pose major democratic challenges by hindering users' ability to make informed and autonomous decisions. By manipulating choices and behaviour via interfaces deliberately designed to direct the user's action (such as hidden subscriptions, complicated cancellation options or biased default choices), these practices impact on the transparency of the service, thereby calling into question the degree of information that needs to be given to a user in order to encapacitate them, a central issue in user experience design research. In this case, the reduction in transparency weakens the notion of informed consent, which is essential in a democracy, where individuals must be able to exercise their rights in full knowledge of the facts. These manipulations also undermine citizens' individual freedoms, by reducing their ability to act in accordance with their true interests. Finally, they blur the boundaries between information and influence, making it difficult to distinguish between personal will and commercial pressure, thereby compromising confidence in digital services and, more broadly, in the democratic institutions that should be protecting consumers.Pour autant, toutes ces interfaces ou éléments d'interface ne sont pas nécessairement conçus dans une stricte optique de manipulation des utilisateurs, beaucoup relevant des habitudes de conception, du respect des critères de performance du design d'expérience utilisateur (utilisabilité, engagement, conversion) ou de l'inspiration d'autres services (pratique de benchmark).

In this sense, there is a lack of tools for comparing the level of persuasion of digital services between different services and identifying areas for improvement in terms of respect for individual freedoms.

To meet these challenges, this thesis project addresses an original field in the classification of dark patterns, by focusing on the ontological analysis of digital interfaces in terms of their persuasiveness. In this sense, the aim is to determine the conditions for the emergence of a dark pattern, based on a conceptual model of the digital interface. We hypothesise that by precisely identifying the components of a digital interface, the relationships between them, their contexts of use and their possible parameters, it will be possible to determine certain component configurations that will facilitate the detection of dark patterns in interfaces. In this sense, the thesis project aims to contribute to the production of tools for highlighting dark patterns in users (for example in the form of a score such as the Nutriscore).