SOCIETÀ ITALIANA DI DIRITTO ED ECONOMIA
cecilia isola (University of Genoa)
Fabrizio Esposito (Nova Law school (Lisbon))
Abstract
‘Dark pattern’. It seems like everyone is concerned by them. But what are they?
Policy-oriented debates are often driven by a perception of the great salience of a certain issue. When this happens, the expression associated with that idea becomes a buzzword and its meaning somewhat uncertain. In recent years, we have seen this phenomenon for example with ‘nudges’. At some point, the impression was that almost all attempts to influence somebody else would be nudge.
In the case of dark patterns, the impact on public discourse is perhaps even more explicit. This is particularly evident in the European Union. First, the European Commission has funded a study entitled ‘Behavioural study on unfair commercial practices in the digital environment - Dark patterns and manipulative personalisation: final report’.1 This study was commissioned specifically to investigate digital market conditions and assess whether the current EU consumer protection regulatory framework was effective in addressing the challenges posed by practices such as dark patterns. Second, the main legal instruments introduced by the previous Commission to organize the EU digital market refer to dark patterns, namely the Digital Services Act, the Digital Markets Act, and the Artificial Intelligence Act.
Against this background, this article seeks to bring some clarity. More precisely, this article combines a systematic literature review with conceptual analysis to, ultimately, justify the following claim: the conceptual framework proposed by the unfair commercial practices directive offers the most solid standpoint for reflecting on dark patterns; dark patterns may be misleading (by action or by omission) and/or aggressive (undue influence, harassment, coercion) and do not necessarily exploit heuristics and biases.
This claim is justified as follows. A systematic literature review is used to map the literature around three topics: definitions of ‘dark pattern’; examples of dark pattern; qualification of dark patterns under EU law. The findings regarding each of these topics is then further developed.
Definitions of ‘dark pattern’ are connected with the broader reflection prompted by digital markets on the idea of manipulation. At the same time, our SLR presents an updated list of dark patterns. The list demonstrates that – contrary to a widespread opinion emerging from the literature – not all dark patterns are connected to heuristics and/or biases.
The SLR also shows essentially unanimous consent among EU legal scholars regarding the qualification of dark patterns as unfair commercial practices within the meaning of Directive 2009/25 (Unfair Commercial Practices Directive; UCPD). Building on this finding, the article investigates the hypothesis that dark patterns are best classified using the UCPD framework.
Accordingly, the article classified dark patterns under the UCPD categories: misleading action; misleading omission; harassment; undue influence; coercion.
A comparison between the resulting classification and the classifications by the Behavioural Study (relying in particular on Luguri et al. 2021) and Mathur et ali. 2021 suggests that the UCPD offers a better classificatory framework.
This conclusion is further strengthened by combining the results of our SLR with those of a different one on dark patterns (with a slightly different data collection protocol and different research questions). The combination confirms that the concerns associated with dark patterns are the same UCPD is meant to address. This finding confirms the connection between dark patterns and the UCPD.
The Article is divided in two parts. Part I is dedicated to the Systematic Literature Review (SLR). It details the methodology employed, including the use of the PRISMA protocol, and presents the results of the review. Part I provides an overview of the sources examined and the key themes emerging from the existing literature on dark patterns.
Part II integrates the results of the SLR with the regulatory framework of the UCPD. To propose a clear and UCPD-inspired classification (UCPC) of dark patterns that will be shown to be superior to the one proposed by the Behavioural Study. The superiority is two-fold. In absolute terms, the UCPC is conceptually more convincing. Moreover, the UCPC establishes better grounds for interdisciplinary dialogue between social and interface design experts and the law community, thereby reducing linguistic barriers to effective enforcement.
The article makes five contributions. First, the article offers much-needed clarity regarding the meaning of ‘dark pattern’.
´Second, it provides an improved classification of the dark patterns identified in the literature. Third, by connecting dark patterns to a well-established EU legal framework, the article simplifies enforcement in the European Union. At the same time, present and future anti-dark pattern legislation in Europe and beyond can be better drafted and interpreted by relying on this classification. Fourth, the comparison between the UCPD classification of dark patterns and the Cognitive Bias Codex helps construct convincing theories of harm, thereby strengthening enforcement further.
Finally, from a methodological point of view, the article illustrates the often underestimated analytical value of existing legal structures. Thus, the article suggests that social scientists pay more attention to legal dogmatics in the construction of their theoretical frameworks.