Deconstructing the right to privacy considering the impact of fashion recommender systems on an individual’s autonomy and identity

Onitiu, Daria (2021) Deconstructing the right to privacy considering the impact of fashion recommender systems on an individual’s autonomy and identity. Doctoral thesis, Northumbria University.

Preview

Text (Doctoral thesis) onitiu.daria_phd(17024841).pdf - Submitted Version Download (3MB) | Preview

Abstract

Computing ‘fashion’ into a system of algorithms that personalise an individual’s shopping journey is not without risks to the way we express, assess, and develop aspects of our identity. This study uses an interdisciplinary research approach to examine how an individual’s interaction with algorithms in the fashion domain shapes our understanding of an individual’s privacy, autonomy, and identity. Using fashion theory and psychology, I make two contributions to the meaning of privacy to protect notions of identity and autonomy, and develop a more nuanced perspective on this concept using ‘fashion identity’. One, a more varied outlook on privacy allows us to examine how algorithmic constructions impose inherent reductions on individual sense-making in developing and reinventing personal fashion choices. A “right to not be reduced” allows us to focus on the individual’s practice of identity and choice with regard to the algorithmic entities incorporating imperfect semblances on the personal and social aspects of fashion. Second, I submit that we need a new perspective on the right to privacy to address the risks of algorithmic personalisation systems in fashion. There are gaps in the law regarding capturing the impact of algorithmic personalisation systems on an individual’s inference of knowledge about fashion, as well as the associations of fashion applied to individual circumstances. Focusing on the case law of the European Court of Human Rights (ECtHR) and the General Data Protection Regulation (GDPR), as well as aspects of EU non-discrimination and consumer law, I underline that we need to develop a proactive approach to the right to privacy entailing the incorporation of new values. I define these values to include an individual’s perception and self-relationality, describing the impact of algorithmic personalisation systems on an individual’s inference of knowledge about fashion, as well as the associations of fashion applied to individual circumstances.
The study concludes with recommendations regarding the use of AI techniques in fashion using an international human rights approach. I argue that the “right to not be reduced” requires new interpretative guidance informing international human rights standards, including Article 17 of the International Covenant on Civil and Political Rights (ICCPR). Moreover, I consider that the “right to not be reduced” requires us to consider novel choices that inform the design and deployment of algorithmic personalisation systems in fashion, considering the UN Guiding Principles on Business and Human Rights and the EU Commission’s Proposal for an AI Act.

Actions (login required)

View Item

Downloads