Algorithmic Profiling and the Data Protection Regulation

Abstract

The thesis provides an overview of the transparency obligations within the European General Data Protection Regulation (GDPR), which social media are subject to, when using recommender systems. The thesis casts light upon the transparency principal envisaged in the GDPR article 5 by researching what is deemed necessary to be informed by the social media. In relation hereto central articles are included; such as GDPR article 13 and 15. Additionally, the thesis wishes to clarify whether any transparency obligations may be deduced from the GDPR article 24 given the risks associated with the processing of data. Finally, with the exemplification of Facebook and Instagram, the thesis concludes whether the GDPR allows such use of recommender systems. Ultimately, the GDPR creates a legislative framework that de jure does not prohibit the use of recommender systems but would de facto be a hindrance for the use of recommender systems. Although the current regulation is unclear and requires further clarity from EU judicial institutions.

In the economic segment of the analyses, the thesis uses game theory to illustrate and analyse the incentives behind usage of recommender systems in the business segment of social media. By illustrating optimal strategies on both the user and social media level. The latter part of the economic analyses draws it theoretical background from the principal-agent theory. Hereby introducing the adverse selection problem of users preferring privacy over personalised content and how the regulation may influence the choices of both users and social media. The segment concludes that social media no matter the possible sanction of breaching regulation would always rather pay the fine, then attempt to be compliant with current regulation.

Finally, the integrated chapter uses the results from the economic segment in relation to the newly established Digital Service Act (DSA), which is yet to be applied. The integrated analysis focuses on the possible effects of the DSA and how it may affect the current legal environment and the choices made by social media. Additionally, Schumpeter’s theory is introduced to analyse whether the DSA will incentives social media and the market in general to invest in and innovate the AI. The chapter concludes that the DSA would solve the intrinsic adverse selection problem. Yet the DSA would impair some incentives for investment in innovative project. On the other hand, it would also incentivise social media and entrepreneurs to redirect their investment and innovative efforts towards developing Explainable AI solutions.