November 10-11 2021, Dapde successfully hosted its first digital conference on Dark Patterns. The aim of the event was to look at the topic from different angles and to discuss existing and possible future answers to current legal frameworks as well as technical solutions. To illuminate this topic, the conference consisted of top-class panelists from industry, science, and consumer protection organizations.
Introduction
After an address of welcome by project leaders Prof. Mario Martini (German Research Institute for Public Administration Speyer) and Prof. Michael Gertz (University Heidelberg), Prof. Christian Kastrop, State Secretary BMJV (German Federal Ministry of Justice and Consumer Protection) introduced the topic of Dark Patterns by referring to ongoing legislative initiatives addressing the issue. The Digital Services Act fully recognizes the danger of Dark Patterns and the necessity for a ban which is promising as a European uniform regulation would offer high effectiveness. Similarly, the Artificial Intelligence Act includes a ban on Dark Patterns and aims at achieving harmonized rules for Artificial Intelligence. Prof. Kastrop emphasized the psychological and economic harm caused by Dark Patterns and called for an examination of the General Data Protection Regulation and German Telemedia Act. Here, Prof. Kastrop called for consumer protection 2.0 which includes empowerment of civil society through collective rather than individual action. Prof. Kastrop concluded by stating that consumer protection must be taken to a new level in which new technologies are at the forefront of ensuring rights for consumers.
Panel 1: What´s the problem? Dark Patterns and Online Manipulation
“Before you regulate, you need to define what you need to regulate.” Colin M. Gray (Purdue University), Susanne Hahn (University of Düsseldorf), and Kat Zhou (Spotify) opened the first panel by presenting Dark Patterns as a multidimensional and complex problem. The discussion, moderated by Hannah Ruschemeier (CAIS), revealed that the current model of the homo oeconomicus does not represent the reality of rational human behavior, which poses a challenge to the ethical and normative dimensions of a user-friendly digital design. At the current technological scale, the user’s agency over their own actions is limited. Therefore, designers and technologists might have to reconsider users’ awareness and ability to act. To harmonize the commercial interest of consumers and buyers, institutionalized transparency was proposed. Individual responsibility could be replaced by an institutional collective solution: Instead of solely burdening the users with an imbalanced decision-making process, institutions could offer process and product certification mechanisms to create transparency and trust in the decision-making process.
Panel 2: Will existing rules save us? Dark Patterns and Data Protection
The second panel, moderated by Christian Drews (Dapde), turned to the data protection and privacy context. Cookie banners that nudge users into giving their consent to extensive data processing are one notorious example that the panel focused on. Prof. Frederik Zuiderveen Borgesius (University of Nijmegen) initiated the discussion by giving some insight on the legal framework revolving around Dark Patterns. He was followed by Estelle Hary (CNIL/EDPB), who introduced the perspective of national data protection authorities. Although the awareness of Dark Patterns and businesses´ GDPR compliance is currently growing, there is a lack of expert understanding on user experience design. Simultaneously, the design community is still lacking the necessary legal knowledge. Ms Hary called for a bridging of this gap. Max Schrems (NOYB) shed light on the phenomenon of “consent optimization” whilst considering the various types of dark patterns. According to Mr Schrems, most dark patterns can be captured by the fairness and transparency principles of the GDPR. Specifically, Article 12 indicates that consent shall be as easily declined as it may be given.
The following discussion focused on whether current substantive law is sufficient or whether new rules are needed. Schrems and Borgesius stressed the lack of operationalization when applying the GDPR and ePrivacy Directive to Dark Patterns. The panel tended to agree that even if the legal framework may be sufficient, a major problem lies within enforcement deficits. During the discussion, many proposals were raised on how to improve the existing legal framework. Regulators could specify the meaning of “fairness” by making existing rules more specific (Borgesius). Exemplary of a dynamic, user-friendly, and technology-neutral approach is the binding anti-tracking browser signal as proposed by the European Parliament for the ePrivacy-Regulation. Additionally, companies could be required to justify their design choices (Hary). Further, collective actions for privacy cases could pose a solution for small but frequent individual damages in consumer and data protection law. The discussion concludes that the spirit of the GDPR, encompassing the demand for clear information, transparency, and control for data subjects, is violated by dark patterns. Whilst a call for stricter transparency requirements may seem intuitive, this could continue to burden individuals with the responsibility to respond to dark patterns. Therefore, as stated in the beginning, collective over individual action seems to be necessary to stop unfair commercial practices.
Panel 3: Focusing on the most vulnerable - Dark Patterns targeting Children/Teenagers in Apps and Games
The third panel, moderated by Paul Seeliger (Dapde), created an interesting analogy to Dark Patterns: Game design techniques. Just like Dark Patterns, mechanisms in contemporary game design require a critical examination, especially considering its young audience.
Prof. Jose Zagal (University of Utah) introduced Dark Patterns in apps and games by giving the perspective of game designers. Quoting Game Designer Paul Czege, Zagal characterized game design as “the art of getting people to behave as players without your instructions” and therefore inherently manipulative. Essentially, game design is a form of seduction, aiming at keeping people engaged. This engagement is followed by the “paradox of games” describing how in games, something “bad” can be “good”, standing in contrast to the interest of the consumer. Whilst manipulation techniques are natural to the essence of games themselves, Dark Patterns serve a different purpose. The design of Dark Patterns leads users to act in a certain way which is contrary to their interests, meaning that the design power unilaterally benefits the interest of the creator. This is different from the mutually beneficial manipulation techniques in games. Zagal concluded by presenting the development of the game industry, which has always been experimenting with design and monetizing these designs.
Prof. Angela Campbell (Georgetown University) shed light on how manipulative techniques in games target children by presenting the Barbie Dreamhouse Adventures Application as a case study: As children cannot understand the implications of privacy matters when playing games, the manipulative techniques of the game’s storyline are very effective. Children develop a para-social relationship with the characters in the game which makes them very receptive to purchase requests expressed by the characters themselves. Therefore, the question appears whether more financial control in games is needed: The discussion revealed that a weighing of interests is crucial to determine whether the parents or rather game designers themselves should bear responsibility, implying a rather paternalistic approach.
Panel 4: Will new rules save us? Dark Patterns in E-Commerce
Day 2 started with Panel 4 on Dark Patterns in E-Commerce, moderated by Inken Kramme (Dapde) featuring Anne-Jel Hoelen (Authority on Consumers and Markets of the Netherlands), Miika Blinn (Federation of German Consumer Organisations), and Prof. Hans Micklitz (European University Institute Florence) as speakers. The focal point of the panel was the pending EU-regulation regarding Dark Patterns and consumer protection, especially the Digital Markets Act, the Digital Services Act, and the Artificial Intelligence Act.
The discussion revealed that the current drafts of the DMA and DSA do not contain any consumer protection regulations and it is highly improbable that this will be added in the final drafts. Instead of posing new rules for consumer protection, it would be efficient to state that general consumer protection still applies in the digital world. Despite the Unfair Commercial Practices Directive, current regulations are considered too narrow, impeding proof and enforcement. However, the outcome of the DSA and DMA will be noticeable for the consumer in other aspects, e.g., with the installation of apps programmed by other app stores other than the one pre-installed on your phone. The panel discussion revealed that the Artificial Intelligence Act serves as a positive example in constituting a productive step in the diffusion of Dark Patterns, introducing the concept of “digital vulnerability”. Yet, the panelists were not to conclude that the newly proposed regulations will be of great effect, leaving the already existing Unfair Commercial Practices Directive as the most promising regulation regarding consumers and Dark Patterns.
Panel 5: Will Technology save us? On recognizing Dark Patterns automatically and preventing them
The conference’s last panel, moderated by Prof. Gertz, shifted its view from the legal to a more technical perspective. Midas Nouwens (Aarhus University), Marija Slavkovik (University of Bergen), and Harshvardhan Pandit (Trinity College Dublin) discussed whether apps and other technical approaches might be a solution for digital manipulation. While all panelists acknowledged the important role of automatized scraping for Dark Patterns or cookie administration tools, they also agreed that technology alone will not be sufficient to protect individual autonomy.
As an introduction, Nouwens stressed that the problem of Dark Patterns cannot be solved through Artificial Intelligence alone: In fact, regulations and media attention around Dark Patterns are immensely important in creating direct effects. Based on Nouwens empirical observations in Denmark, especially media attention triggers an effective governmental response. Following this, Slavkovik extended the conversation by thematizing machine ethics: She ascribes special importance to the definition of dark patterns for machine learning algorithms; describing the experience of users rather than its specific design elements. Pandit extended on the importance of semantics for legally defining a dark pattern: It is crucial to find the correct elements and vocabulary to create an interdisciplinary understanding of a dark pattern´s nature. The following discussion revealed the difficulty of defining a dark pattern, as ultimately not only the effect of a dark pattern's perception is very subjective, but also the existence of having to make a choice in the first place could be seen as manipulative. Concluding, although many issues are yet to be solved, the panel agreed that technology can be a valid remedy against dark patterns considering that Dark Patterns are first and foremost a technological phenomenon.
by Katharina Buchsbaum
