Nowhere to hide: an ethical evaluation of how big data aggregation violates privacy (and what we should do about it)

Date
2023-03
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH ABSTRACT: Data aggregation involves merging different kinds of data into a single dataset and analysing this data together to retrieve patterns that might exist between this data and in so doing reveals novel information on the subjects of the data. Analysing aggregated data is referred to as aggregation-knowledge discovery in databases (A-KDD). Data aggregation has become popular in Big Data analysis as it allows for more and more complex information to be retrieved from otherwise innocuous data points. This thesis argues that data aggregation can result in unique violations of privacy that can have a negative impact on the individual and social wellbeing of a data subject. To make this argument, time is spent unpacking the concept of privacy and its value in liberal democratic societies. Following this, a description of big data and data aggregation is given. Once these terms are cashed out and understood, A-KDD is shown to result in violations of privacy on two accounts. In cases where the well-being of an individual is potentially decreased due to A-KDD revealing unknown information on a person, it is considered unjustifiable, and thus constitutes a (morally problematic) violation of privacy. In cases where A-KDD attempts to access unknown information without the consent of a data subject, it is also considered unjustifiable and thus constitutes a (morally problematic) violation of privacy. The thesis closes with suggestions on how A-KDD might be regulated to ensure the privacy violations of the practice are mitigated.
AFRIKAANSE OPSOMMING: Datasamevoeging behels die samevoeging van verskillende soorte data tot ’n enkele datastel en die analise van hierdie datastel om patrone binne die data rte vind en sodoende oorheenonbekende inligting oor die datasubjekte te bekom. Daar kan na sodanige analise van saamgevoegde data as samevoegende-kennisontdekking in databasisse verwys word. Datasamevoeging vier hoogty in Grootdata-analise omdat dit die ontdekking van meer en meer komplekse inligting vanuit gegewe datapunte moontlik maak. Ek voer in hierdie tesis aan dat datasamevoeging tot nuwe, unieke vorms van privaatheidskending lei wat ’n besonder negatiewe impak op datasubjekte se individuele en sosiale welstand kan hê. Om hierdie argument te maak, ondersoek ek eers die aard van privaatheid asook die waarde daarvan binne die konteks van ’n liberaal-demokratiese bestel. Hierna beskryf ek grootdata en grootdataanalise. Dan verduidelik ek hoe die proses van samevoegende-kennisontdekking in databasisse op twee maniere tot die skending van privaatheid kan lei. In gevalle waar sensitiewe inligting ontbloot word wat sodoende ’n negatiewe impak op individue se die welstand kan hê, bevind ek dat die praktyk onregverdigbaar is en dus uitloop op moreel-problematiese privaatheidskending. Dieselfde geld vir gevalle waar daar gepoog word om onbekende inligting oor individue sonder hulle toestemming te bekom. Ek sluit af met ’n bespreking van hoe hierdie data-analisepraktyke gereguleer kan word ten einde privaatheidskending te voorkom.
Description
Thesis (MPhil)--Stellenbosch University, 2023.
Keywords
Citation