1

Gintarė Surblytė-Namavičienė, Lecturer in the Faculty of Law at Vilnius University, asks in her comprehensive book on competition and regulation in the data economy, whether artificial intelligence needs a new balance. The cardinal question is whether AI has in fact changed fundamental economic parameters which would demand drastic legal changes. And in the end, she pleads for a fine-tuning of the legal framework, rather than for radical legal changes. How Surblytė-Namavičienė arrives at her conclusion with regard to several different, but yet linked, aspects of regulation (i.e. trade secret law, data protection, competition law, and consumer protection), becomes clear when delving into her intelligible and thought-provoking analysis. The result is a book that is much worth reading.

2

Chapter 2 lays the foundation by introducing the functional characteristics and essence of the digital economy. Surblytė-Namavičienė clarifies upfront what is meant by artificial intelligence and highlights the decisive distinctions between “general” and “narrow” AI and between “strong” and “weak” AI. While general AI still appears as a utopia, she highlights the learning mechanism as distinctive feature of AI. From a standpoint of Luhmann’s systems theory, AI cannot be considered as an ‘autopoietic system’, but at least the increasing use of AI may disrupt the interaction between individuals and therefore the basis of existing systems. [1] For this reason, AI can have a significant social impact. Yet from the perspective of economic theory, Surblytė-Namavičienė regards the data economy not as a “revolution”, but rather as a development which follows classic economic principles. Based on the work of Adam Smith, she highlights the significance of self-interest for the functioning of the data- and algorithm-driven economy, which can also explain the ‘privacy paradox’ in her view. Yet, what has indeed changed is the general importance of the economic role of data, which has dramatically increased over the last years. While Surblytė-Namavičienė regards the regulatory debate on data access and ownership as essential, she criticizes the focus on non-personal data as being too narrow. For this reason, she then goes into more detail about the difference between personal and non-personal data, which forms the basis for the following chapters, in which she addresses the interface between data protection and the other relevant areas of law on several occasions. She concludes by raising the seminal question of how to strike a balance between economic incentives of the undertakings when implementing AI on the one hand and the protection of consumers (also data subjects) on the other hand.

3

Chapter 3 focuses on trade secret protection. The justification for attaching the pole position to this often-overlooked regulatory regime lies in its significance for the data economy: trade secret protection does not depend on intellectual efforts, it may protect datasets as well as algorithms, and it may be extensively applied in practice. Therefore, trade secret protection reaches far beyond IoT-settings, which initially triggered the discussion on the significance of trade secret protection in the digital economy. However, the exact scope and application of rules under Directive (EU) 2016/943 on the Protection of Trade Secrets to the data economy are far from clear. For this reason, Surblytė-Namavičienė performs a comprehensive analysis of the requirements for and legal consequences of trade secret protection, for which she also takes an informative side glance at the protection mechanism in the US. Surblytė-Namavičienė regards data as generally eligible for trade secret protection. She then focuses on personal data as a particular subject matter of trade secret protection. As a consequence, natural tensions occur between the undertaking’s interest to protect such data as a secret on the one hand and data subjects’ rights under the GDPR on the other hand, because claiming such rights under the GDPR may require the undertaking to share the data with the data subject or third parties. This is especially true for the right of access under Article 15 GDPR and the right to data portability under Article 20 GDPR, which Surblytė-Namavičienė takes a meticulous look at. In addition, the right to not be subject of automated decision-making, including profiling (Article 22 GDPR), adds to the tension between data protection and trade secrets, because algorithms that serve automated decision-making may indeed be subject of trade secret protection. Surblytė-Namavičienė then points to the fundamental right to conduct business, which may cover trade secrets, but she concludes that the EU Trade Secret Directive itself does not explain how fundamental rights are to be balanced. This leads to significant uncertainty for the legal application in scenarios where secrecy protection collides with data protection. In future, much remains to be clarified on a case-by-case basis by the courts. Another important aspect is reverse engineering, which is allowed for information protected under the trade secrecy rules. Surblytė-Namavičienė argues that for effectively enabling reverse engineering, it is necessary to refuse trademark protection for functional signs, while contractual restrictions may nevertheless prevent reverse engineering. Although trade secrets undoubtedly play an important role for the data economy, regulating algorithms reaches beyond trade secret law, especially with regard to competition.

4

Chapter 4 therefore deals with competition, the key question being how much ‘rethinking’ of competition law is needed in light of the technical developments of recent years. In particular, Surblytė-Namavičienė puts three issues under the microscope. First, she examines algorithmic price adjustments, which the competition law community started to discuss comparatively early. Regardless of this phenomenon’s actual practical significance, which Surblytė-Namavičienė puts into question, she extensively analyzes the standard on price-fixing and concerted practices under Article 101 TFEU. She illustrates how the CJEU’s E-Turas decision [2] has considerably broadened the scope. This decision leaves us with significant uncertainty and further blurs the line between concerted practices and mere parallel behavior. Surblytė-Namavičienė considers the legal implications of the E-Turas decision as highly relevant for the algorithm-driven economy and warns against overenforcement of EU competition law in this domain. The second issue relates to competition for data traffic. This concerns selective distribution as well as rights relating to datasets. Regarding the latter, Surblytė-Namavičienė reflects on the crucial sui generis right for databases under Directive 96/9/EC, which illustrates the significance of exclusive rights protection from a competition point of view. She accurately highlights the importance of the CJEU’s Ryanair decision [3] for the data economy, according to which merely contractual restrictions to data scraping are valid if the database is not protected under the sui generis right. According to Surblytė-Namavičienė, such contractual restrictions can generate anticompetitive effects and may negatively affect consumers by preventing them from choice. The third issue concerns data access under Article 102 TFEU. [4] Surblytė-Namavičienė argues that for such access claims, the “exceptional circumstances test” [5] from the IMS Health case should not be overestimated, because this case depended on specific facts and appears rather informative regarding its implications for unfair competition. Instead, the CJEU’s Bronner decision, [6] which sets out a “pure” indispensability requirement, would provide the relevant legal standard for claiming access to data on the basis of Article 102 TFEU.

5

Chapter 5 then broadens the view beyond competition law and asks which other regulatory regimes become relevant for the data economy. Here, Surblytė-Namavičienė focuses on the threat of algorithmic manipulation, especially in the fields of personalized services and personalized pricing in the business-to-consumer relationships and with regard to rankings by online platforms. After elaborating on these issues, she identifies a regulatory gap with respect to the protection of consumers and calls for regulation which should ensure transparency and prohibit certain behavior for undertakings. In this regard, she considers the already existing regulation of algorithmic trading of financial instruments as informative. A further aspect for regulation is consumer contract protection, in relation to which Surblytė-Namavičienė pleads for “more robust state control of terms and conditions”. In particular, she highlights the significance and complexity of consent regarding the use of personal data as well as the role of competition law by discussing the infamous Facebook decision of the Bundeskartellamt. [7] Finally, she remains critical with regard to approaches of self-regulation, especially when fundamental rights and privacy are involved, as is often the case with AI-driven markets.

6

These chapters reveal how Surblytė-Namavičienė elaborates on a wide range of topics, which are undoubtedly all highly relevant for the functioning and development of the data and algorithm-driven economy. Of course, they cannot be held as exhaustive, and rather than a holistic picture, the analytical depth and focus on selected issues and the well-considered hinting to important links between regulatory regimes is a particular strength of the book. This work is especially informative for researchers who deal with trade secrets, algorithmic collusion, access to data under competition law, and the competition/data protection interface. Surblytė-Namavičienė refers to classical thinkers (such as Smith, Turing, Arrow, Coase, and Schumpeter), and she explicitly justifies her focuses before spotting respective legal uncertainties, which indeed need more clarification. In substance, one could argue that classic economic theory has been contested on grounds of behavioral economics. In fact, Surblytė-Namavičienė acknowledges the role that psychological effects play in competition, while leaving it to the reader to think about what impact they might have on the found solutions. Overall, AI technology has not changed the underlying economic principles based on which the data economy functions as such (e.g. the economic ingredients of the platform economy were all already known). However, the effects of different forces working together have led to unprecedented situations, which indeed challenge the law. Therefore, one can ask what circumstances would lead to a drastic change and which parameters and contexts are relevant to understand when a change of paradigm is needed for approaches to regulate the data economy.

7

Some significant developments haven taken place after the publication of the book, and they could therefore not be considered. This is true for the German Facebook decisions of the OLG Düsseldorf and the Federal Court of Justice [8] and the recent reform of the German Act Against Restraints of Competition. [9] Also, the book could not take the Commission’s proposals for a Digital Market Act  [10] and a Digital Services Act into account, [11] which in fact address some of the issues Surblytė-Namavičienė elaborates on. Furthermore, the upcoming Data Act (the Commission’s proposal is expected to be published in Spring 2022) aims to address the intersection between trade secrets and data economy [12]. In this regard, it would be wise for the EU legislator to consult Surblytė-Namavičienė’s book.

8

In times when academic writing on digital regulation tends to compete for the most visionary and most revolutionary approach, the route taken by Surblytė-Namavičienė is remarkably critical, prudent, and cautious. This contributes to the attractiveness of the work, as she clearly delineates the potential and limitations of competition law and critically highlights the crucial interfaces between the regulatory regimes. Surblytė-Namavičienė disregards many common assumptions as “speculative”, “overestimated”, “exaggerated”, and “hypothetical”. Rather than claiming that things are, she prefers to say that they might or could. This absence from overhasty generalizations appears like an honest approach that puts, however, the question for empirical evidence and its significance for evidence-based policy making on the table. Here, the book asks the right questions, but answering them in a definite way would require an extensive evaluation of empirical research results, which would surely go beyond the book’s scope. As a consequence, Surblytė-Namavičienė does neither provide speculative answers, nor do her suggestions on how to adjust the legal framework become overly concrete. Rather, the reader gains inspiration and is indeed left with the sensible claim that it is all about the fine-tuning of the legal framework. Surblytė-Namavičienė rightly points to the neuralgic spots and, even more so, urges for timely reforms in this regard. Considering the recently initiated but by far not yet completed legislative actions on the EU level, it appears too early to tell though whether this remains wishful thinking in light of political realities.

by Heiko Richter, MPI München