Page - 42 - in Austrian Law Journal, Volume 1/2019

ALJ 2019 Peter Egger et al 42 made to explicitly address at least some aspects of the use of (machine-learning) algorithms, with Art. 22 of the General Data Protection Regulation (GDPR)26 being the first attempt to do so (see below 2.2.1.). In addition to binding legislation, the EU seems to approach the challenges of the digitisation in general and the use of machine-learning algorithms in particular by publishing (Commission27) policy documents on the one hand (namely the “Communication on AI for Europe”28 and the Communication on “Liabilities for emerging digital technology”29) and by raising awareness on the other hand through initiatives like the Algorithm Awareness Project.30 In addition, the Commission established a High-Level Expert Group on Artificial Intelligence (AI HLEG). This independent expert group agreed on a definition of AI in April 2019 which could potentially become legally binding in the future.31 To sum up: Adequate measures are needed to address the legal challenges of the increasing use of algorithms. As a first step, awareness for the relevant (legal) challenges must be created. Without a doubt, the EU is currently working hard to achieve this goal. Its efforts are being accompanied by initiatives on the domestic32 and the international33 level. In some aspects of the use of AI, however, the EU has moved beyond the task of creating awareness, as legally binding instruments already address specific problems. Two examples of such legislation will be given in the next section. B. EU law provisions addressing the use of algorithms As mentioned above, Art. 22 GDPR can serve as an example of an explicit regulation on the use of algorithms and will be discussed below (2.2.1.). However, the lack of explicit rules does not mean that existing legal provisions – without providing a direct link to algorithms at first sight – cannot be suitable to meet the challenges associated with their use. In subsection 2.2.2., Art. 101 TFEU will be discussed as an example of a provision that seems well suited to address the challenges without a need for further adaption even though it does not explicitly mention algorithms. 26 For more details on this provision, see the following remarks in this paper in chapter 2.2.2. 27 The European Parliament has also expressed its opinion on this matter as well as related issues. Cf. European Parliament, Report with recommendations to the Commission on Civil Law Rules on Robotics (2015/2103 [INL], 2017). 28 COM (2018) 237 final. Cf. also Commission, Coordinated Plan on Artificial Intelligence (Communication), COM (2018) 795 final. 29 Commission, Liability for emerging digital technologies (Communication), SWD (2018) 137 final. 30 Cf. COM, Algorithmic Awareness-Building, Apr. 25, 2018, available at https://ec.europa.eu/digital-single- market/en/algorithmic-awareness-building. 31 High-Level Expert Group on Artificial Intelligence, A definition of Artificial Intelligence: main capabilities and scientific disciplines, Apr. 8, 2019, available at https://ec.europa.eu/digital-single-market/en/news/definition- artificial-intelligence-main-capabilities-and-scientific-disciplines; Reinisch, Künstliche Intelligenz - Haftungsfragen 4.0 und weitere zivilrechtliche Überlegungen zu autonomen Systemen, 37 ÖSTERREICHISCHE JURISTENZEITUNG 298 (2019). 32 According to initial reports, an Austrian robotics and AI strategy is planned for the summer of 2019 (cf. APA, Österreichische Robotik- und KI-Strategie für Sommer 2019 geplant, PRESS RELEASE, NOV. 22, 2018, available at science.apa.at/rubrik/politik_und_wirtschaft/Oesterreichische_Robotik-_und_KI- Strategie_fuer_Sommer_2019_geplant/SCI_20181122_SCI40111351045600836. 33 In May 2019, as an important step in this direction, the OECD (with 36 Member States) formally adopted the first set of intergovernmental policy guidelines on Artificial Intelligence that were also endorsed by six more states. Cf. OECD, Forty-two countries adopt new OECD Principles on Artificial Intelligence, May 11, 2019, available at http://www.oecd.org/science/forty-two-countries-adopt-new-oecd-principles-on-artificial-intelligence.htm.