City Research Online

Abstract

Mainstream privacy theory’s traditionally siloed nature gives rise to difficulty in rationalising the intuitive sense that deepfakes threaten privacy, as distinct from other interests. The chapter examines this difficulty in detail, by drawing a distinction between the ‘process’ and ‘product’ phases of deepfake creation, and by comparing these phases with an analogous human activity – the painter who imaginatively constructs a nude portrait from memory. This comparative framework illuminates how process and product operate differently in human and AI‑generated artefacts. It then proposes that a method of ‘triangulation’ can be used to salvage mainstream privacy theory. By locating a degree of consensus between various theories in respect of the conclusions they would reach on the threat deepfakes pose for privacy, we can rationalise the conclusion that deepfakes do indeed pose privacy problems without the need for consensus on the principles that underpin that conclusion. In this way, notwithstanding the traditionally siloed nature of privacy scholarship, the triangulated consensus demonstrates that privacy concerns arise at the product stage of deepfake creation, though not at the equivalent stage of creating the nude painting, thereby providing a principled basis for treating deepfakes themselves as distinct, privacy‑bearing artefacts.

Publication Type:	Book Section
Additional Information:	This is an Accepted Manuscript of a book chapter to be published by Routledge. It will be available online at http://www.routledge.com/
Subjects:	K Law > K Law (General) Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Departments:	The City Law School The City Law School > Academic Programmes
SWORD Depositor:	Symplectic Administrator

Export

Downloads