INTRODUCTION

• 1. It is unthinkable to imagine a world without computers and the Internet. Pretty soon, it will be unimaginable to live in a world without robots and AI. Increased automation is currently becoming part of our lives and is not only limited to the industry sector. It has also penetrated the services sector and soon it will become prominent in our everyday lives. As it appears from different chapters in this book, increased automation presents many opportunities to different elements of our society but also many challenges. This is not different in the framework of taxation.

• 2. In the current state of play, a specific tax on robots does not exist in Belgium. Rather, the profits of corporations generated by robots are subject to tax according to the general corporate income tax (CIT) rules. In this respect, most states treat robots owned by corporations as assets similar to machines or computers which form part of the process of the production of goods or the performance of services. However, there are some governments that have issued proposals or implemented legislation aiming at taxing automation because it is perceived as presenting a threat to the labour market. For example, South Korea has reduced tax incentives for corporations using robots3 and New York City mayor Bill de Blasio has proposed a specific tax on large corporations that use robots to the detriment of labour forces. In this regard, it should be observed that, even in the few cases in which a tax on automation has been proposed/implemented, there is not yet a comprehensive and generally agreed system of taxation of robots amongst countries. In Belgium, the profits derived by robots are taxed together with the other profits of a corporation according to the general CIT rules. However, certain tax incentives apply for companies investing in robots/AI.

• 3. Although robots are as such not yet being taxed, governments have recognised that the use of robots/AI (and the digitalisation of the economy in general) has changed business models in such a way that corporations can scale up without mass and as such do no longer need the presence of premises or personnel in a particular country in order to conduct their business there.

INTRODUCTION

• 1. At the shareholders meeting in 2016, Elon Musk presented the Tesla factory as ‘the machine that makes the machine’. He referred to this machine-making machine in the context of Tesla's manufacturing plant that would not need much human involvement. The goal was to produce cars by using smart robots and with as few employees as possible in the assembly line. However, the production of these cars incurred serious delays. In 2018, Elon Musk had to admit that he had relied too much on the capabilities of robots and that ‘humans are underrated’. Another example of underestimating human input in robotised work environments is the robotic production of Adidas shoes. The company had to shut down these factories after two years because the robots could not meet the requirements of the flexible production rate. It was more difficult to retrain than to rely on human workers instead.

• 2. These examples express the idea of the so-called ‘Dark Factory’, which refers to manufacturing plants using intelligent machines and systems that require hardly any involvement of human operators. This would lead to a factory that can operate in the dark without any lights. However, the examples mentioned above show that human workers are still needed in these robotised work environments. This is taken as guiding principle throughout this entire chapter. When it comes to robots that replace human workers, this chapter argues that fully automated working environments cannot be created with the current technological capabilities of robots. On the contrary, robots will perform only a part of the tasks of human workers. Therefore, human workers and robots will have to work more closely together. This will have an impact on the relationship between them. As will be discussed below, this evolving humanrobot relationship will change the working conditions for workers and will affect the employment relationship.

• 3. This chapter does not disregard the fact that robots have been part of manufacturing plants for several decades. Nevertheless, it will be argued that the view of the robot in the workplace as a huge robot arm in the assembly line will soon become outdated or might already be outdated.

INTRODUCTION

• 1. In this chapter we will examine the hypothesis of AI-driven automation, and the resulting technological unemployment and its impact on social security law. After briefly explaining, by way of several examples, what AI-driven automation entails (part 1), this chapter introduces the hypothesis of technological unemployment as a starting point for the analysis (part 2). Assuming this hypothesis is valid, this chapter then discusses its impact on social security law (part 3). In doing so, a distinction is made between employment as a structural feature of the design of social security legislation, and employment as the legitimation for the rights-based character of social security entitlements. Finally, this chapter concludes with the key message that this technological unemployment hypothesis demonstrates the vulnerability of our social security system to radical societal changes caused by technological advancement (part 4).

• 2. At the 2018 Google I/O conference Google CEO Sundar Pichai revealed Duplex, which is an extension of the Google Assistant application. It allows for the automation of conversations, with the presentation showcasing how the virtual assistant can autonomously complete calls such as booking a hairdresser's appointment or making reservations for a restaurant. It did so with notable ease, even if, when making reservations, the human on the other end of the line adds a factor of complexity by misunderstanding what is said. Remarkably, this virtual assistant makes the calls in a surprisingly human-like manner, incorporating several elements in its speech that make it difficult to distinguish from a real human, such as using filler words (‘um’ or ‘hmm’), adopting a human-like intonation, and introducing more response latency to make the conversation feel more natural. Google Duplex is but one of several examples of the recent popularisation of new technologies that involve natural language understanding, others being Amazon's Lex, IBM's Watson, or Apple's Siri. The commercial interest in these technologies is considerable, since they mean a new leap forward in automating tasks such as translation, content analysis, or reasoning. Customer service, an integral part of any organisation's operation, for example, represents a commercial area in which substantial efficiency gains could be made through the existence of intelligent chat bots that never tire, forget, or make errors, and are considerably cheaper than their human counterparts.

INTRODUCTION

• 1. Civil procedure is not an area of law known for its innovative disposition. The roots of civil procedural law for most of the European jurisdictions can be found in the Romano-canonical procedure and this common foundation has seen little profound change since.

This long history does not mean that civil procedure is insulated from innovation. On the contrary, its sometimes arcane and paper-driven methods and procedures leave much room for improvement, an endeavour in which recent changes such as the ‘Potpourri’ legislation in Belgium have only been incrementally successful. Perhaps a shock treatment is necessary for fundamental change within this area of law, and such a shock is what AI aspires to be.

• 2. Artificially intelligent civil procedure goes beyond the mere digitisation of procedure, promising rather a digitised justice in which the daily work of legal actors is to a large extent assisted by machines, changing the nature of judicial decision-making as we know it today. A thorough examination of the possible impact of AI on civil procedure would, at this point in time, largely amount to conjecture. Therefore, this chapter focuses on one phenomenon of which the shape is already discernible, even though successful applications within the area of civil procedure are still few and far between. This phenomenon is quantitative or algorithmic legal prediction (henceforth QLP), colloquially known as predictive justice. As a book chapter of limited length does not lend itself to an in-depth study of this phenomenon, this chapter will only scrape the surface of some topics. The interested reader will find more information in the referenced material. Even though this edited volume focuses on Belgium, this chapter will only briefly touch upon this jurisdiction given the general absence of Belgian initiatives in this field. Rather, the hurdles Belgium needs to overcome in order to facilitate the AI revolution in its civil procedure will be outlined in the final part of this chapter.

• 3. This chapter provides a brief introduction to legal analytics and quantitative legal prediction (part 2), followed by a short overview of some existing use cases of QLP (part 3). Potential use cases and their advantages are discussed (part 4), whilst the challenges that QLP poses are highlighted as well (part 5).

Artificial intelligence (AI) is becoming increasingly more prevalent in our daily social and professional lives. AI can be of benefit to a wide range of sectors such as healthcare, energy consumption, climate change and financial risk management. AI can also help to detect cybersecurity threats and fraud as well as enable law enforcement authorities to fight crime more efficiently. AI systems are more accurate and efficient than humans because they are faster and can better process information. They can perform many tasks ‘better’ than their human counterparts. Companies from various economic sectors already rely on AI applications to decrease costs, generate revenue, enhance product quality and improve competitiveness. AI systems and robots can also have advantages for the specific sector in which they are to be used. Take the example of autonomous vehicles. Transport will become more time-efficient with autonomous car technology. Self-driving cars will also enable people currently facing restrictions for operating a vehicle – such as the elderly, minors or disabled people – to fully and independently participate in traffic. Traffic will become safer as well. The number of accidents will decrease as computers are generally much better drivers than humans.

At the same time, however, the introduction of AI systems and robots will present many challenges. These will only become more acute in light of the predicted explosive growth of the robotics industry over the next decade. AI has implications for various facets of our society. Some even predict that AI systems can completely eradicate humanity in the long run. There are also several important ethical issues associated with (programming and using) AI systems. The commercialisation of AI will pose several challenges from a legal and regulatory point of view as well.

In this comprehensive book, scholars from various legal disciplines critically examine how AI systems may have an impact on Belgian law. While specific topics of Belgian private and public law are thoroughly addressed, the book also provides a general overview of a number of regulatory and ethical AI evolutions and tendencies in the European Union. The book additionally explains basic AI-related concepts such as machine learning, robots, Internet of Things and expert systems.

INTRODUCTION

• 1. AI and IoT have been rolled out with spectacular speed in an increasing number of areas. It is transforming the way in which society and the economy operate. As a late adopter, the insurance industry is also starting to embrace the digital transformation. Insurance companies are exchanging traditional business models for new paradigms such as ‘Connected Insurance’, ‘Usagebased Insurance’ (‘UBI’), and ‘Smart Underwriting’. In the same vein, consumers of insurance products and services are seeking on-demand, personalised experiences. The types of risks being covered are also evolving, presenting fresh business opportunities for insurers to reimagine their business by building a digital-enabled core to deliver sustainable operational efficiencies. The insurance industry is adopting a wide variety of digital technologies to meet these ever-changing demands. These technologies will help insurers to reduce costs, optimise business processes, launch new innovative product offerings and improve customer experience. For all its potential benefits, it bears no doubt that this digital transformation gives rise to legal challenges.

  2. This chapter will elaborate on one specific technological development within the insurance sector, namely the use of telematics in policy underwriting. Underwriting involves measuring risk exposure and determining the premium that needs to be charged to insure that risk. The main technological development which will be the subject of this chapter is UBI. In the context of this chapter, UBI is a type of motor vehicle liability insurance (further referred to as ‘vehicle insurance’) that tracks driving habits and driving behaviour. Basically, UBI means that a driver's behaviour and habits are monitored directly while the person drives, allowing insurers to more closely align driving behaviour and habits with premium rates. Oftentimes, UBI is powered by in-vehicle telecommunication devices (also referred to as ‘telematics’) which are either integrated by the vehicle manufacturers, self-installed using a plug-in device or available through mobile applications.

  3. This chapter will introduce policy underwriting in Belgian insurance in general terms, with a specific focus on the underwriting process in vehicle insurance (part 2). Afterwards, the concept of UBI will be elaborated upon (part 3).

INTRODUCTION

• 1. In a way, it is currently ‘the best of times’. While some aspects of our present-day society could definitely be characterised as ‘the worst of times’, technological innovations have revolutionised our way of living and have changed many things for the better. It is simultaneously the epoch of belief and incredulity, where we revel at machines that are seemingly able to create artistic works and, even, invent. The interface of AI and intellectual property (IP) law is readily apparent and the possible topics of this chapter are manifold. This chapter provides a bird’s-eye view of the current status of research into some of the relevant legal issues from a primarily EU, civil law perspective. The present extensive introduction further delineates the topic (part 1). It includes a brief analysis of the possible uses of AI as a tool in the IP sector as well a detour along the essentials of IP law. The two-pronged substantive body of the chapter analyses the main legal challenges raised by AI – primarily machine learning AI – in an IP context. First, focus lies with the IP protection of AI technology under patent law and copyright law (part 2). Subsequently, attention shifts to the IP protection of output generated through or by an AI system (part 3). Could such output be susceptible to IP protection under copyright and/or patent law? If this is not the case, should it, and, if so, who could or should own the rights, and what should the modalities of protection be? Without providing the final answer to these questions, the conclusions of this chapter (part 4) caution against unreservedly tearing down the foundations of IP protection for the mere sake of additional incentive creation.

• 2. AI is a multidimensional, evolving concept with different possible definitions, as explained in chapter 1, part 2 of this book. In this chapter, we will look at AI as the science and engineering of ‘making intelligent machines’, of ‘making computers do things that require intelligence when done by humans’. However, AI algorithms do not function in the same way as the human brain does – in other words, computers do not ‘think’ as we humans do.

INTRODUCTION

• 1. According to Isaac Asimov's 1942 sci-fi story Runaround, the ‘First Law of Robotics’ prescribes that ‘a robot may not injure a human being or, through inaction, allow a human being to come to harm’. The Second law states that ‘a robot must obey the orders given it by human beings except where such orders would conflict with the First Law’. Technology has come a long way since the heyday of sci-fi books and films about robot overlords and robotised weaponry. And while science-fiction literature and Hollywood have surely fuelled the imagination of the general audience, our civilisation is nowhere near robot domination yet. Nevertheless, technological developments in the methods and means of warfare brew at the horizon like an ever more ominous cloud, and the question of whether humanity needs its own Laws of Robotics becomes increasingly prominent.

• 2. As this book amply shows, the potential of robots and AI in law is tremendous. It certainly seems that in many fields of human activity, this great potential is of a largely beneficial kind. Efficiency, autonomy and an unmatched computational power promise sizeable gains in an array of legal domains. In international law, and more particularly in the realm of the conduct of hostilities, there is not as much space for unbridled optimism. This chapter will attempt to shed some light on important questions of international law when dealing with robots. First, the reader is introduced to the basic concepts of international humanitarian law and several prima facie concerns regarding their relationship to LAWs, in particular the principles of distinction and the prohibition of superfluous injury or unnecessary suffering (part 2). We then explore the legal aspects of LAWs relating to two themes: the authority awarded to machines and automated decision-making processes on wounding and/or killing humans in an armed conflict, as well as the processes and procedural safeguards behind targeting and engagement choices (part 3). This is followed by a briefing on current applications of LAWs and their foreseeable developments, with a particular focus on the US and China as the two military actors most advanced in developing such technology (part 4).

INTRODUCTION

• 1. AI has the power to transform our societies and to have a profound impact across various societal domains. Consequently, it has sparked a debate about the principles and values that should guide its development and use. The concerns include loss of jobs, misuse, discrimination and so on. Against this background, we have in the past years witnessed an ever richer debate on how AI ought and ought not to be (used). Participants in this debate rely not only (and perhaps some would say too little) on the law, that is what is (il-)legal, to provide an answer to this question.

In this chapter, we would like to provide the overarching ethical and legal framework underlying this debate by focusing on three aspects. To fully understand the impact of AI on the law (cf. subsequent chapters of this book), we present the main ethical challenges that AI raises and the principles that have been adopted to tackle them (part 2). We then touch upon a common feature of the policy documents and ethics guidelines on AI that have been published so far: there is a need for some form of regulation. Therefore, before we move on to form and substance of a possible regulation, we first answer some fundamental questions, namely: what is regulation and what, when and how exactly should we regulate (part 3)? Finally, we provide an overview of AI governance in some of the jurisdictions that have already taken policy actions on AI. As this domain is evolving fast, we do not aim to give a comprehensive overview. Rather, we intend to provide a bird's eye view of the direction in which the ethical and legal framework on AI might evolve the coming years (part 4). We conclude with summarising the main findings (part 5).

ETHICS OF AI

• 2. Ethics, which can be described as the study of what is good and bad moral behaviour, has been relied on to the same extent as – and perhaps even more heavily than – the law in the debate on what AI ought (not) to be or do. Although this reliance on ethics has been criticised by some as ‘ethics-washing’, approaching AI from both the perspective of ethics and the law has its merits..

INTRODUCTION

• 1. In this chapter, I will first discuss the rise of robotics and AI in the healthcare sector and the concern of some scholars that this may lead to a dehumanisation of the physician-patient relationship (part 2). I will then elaborate on four potential existing legal safeguards against such dehumanisation: the fact that only qualified persons are allowed to provide healthcare (part 3) and the resulting liability of the physician if things go wrong (part 4); the right of the patient to receive information about his/her health condition and to give his/her prior informed consent under the Belgian Law on Patient Rights (part 5), and finally transparency and informed consent under the General Data Protection Regulation (GDPR) (part 6). I will conclude with an overview (part 7).

THE RISE OF ROBOTICS AND AI TO DEAL WITH INCREASING DEMANDS IN THE HEALTHCARE SECTOR

• 2. A recent publication commissioned by the European Parliament states that the health sector is facing increasing demands on services brought on by issues such as an ageing population, an increase of chronic diseases, budgetary constraints, and a shortage of qualified workers. Developments in the field of robotics and AI can provide countless opportunities for addressing these challenges, resulting in necessary and significant cost and time savings. These efficiency benefits are the result of the fact the work is done more efficiently, more quickly and at a lower cost than a human actor could do it. According to the same study, the application of robotics and AI could lead to improvements in fields such as medical diagnosis, surgical intervention, prevention and treatment of diseases, and support for rehabilitation and longterm care. They could also contribute to more effective and automated work management processes, while offering continuous training for healthcare workers. It is estimated that the market for AI in healthcare will reach around $6,6 billion by 2021 and $8 billion by 2022, with significant cost savings for healthcare systems. According to a recent French study, the health sector is internationally the second most impacted sector by robotics and AI after the telecommunications and technologies sector, but preceding the financial services and automotive sector.

INTRODUCTION

• 1. An innovation is the application of a new or significantly improved product (good or service) or process; a new marketing method; or a new organisational method in business practices, workplace organisation, or external relations. It involves the realisation that products, services, means of production, marketing strategies, delivery methods, and business structures do not take a fixed form but rather are subject to change, either incremental or radical. This chapter concerns innovation policy and what governments ought to do to secure the process by which companies inject novelty into the market. Therefore, it deals with different legal rules for commercial innovations. It does not deal with the underproduction of basic research within an economy or, for instance, the task of the state to produce basic research. Our economies have been engaging with innovation at a rapid pace, and research now revolves around the question of how the tax code can streamline and promote this evolution. Governments have adopted this goal: to ‘promote innovation, encourage the development of new technologies and increase the fund of human knowledge’. Tax law is seen as one of the main toolkits by lawyers, economists, and policy-makers. National economies are competing internationally for the price of the most innovative economy, and our tax codes have been subject to fierce innovation too. In a political environment where governments are pushed to ‘create growth’, favourable R&D treatment has now turned into the panacea of innovation policy. Favourable R&D treatment, such as the allocations of R&D to foreign income, the R&D tax credit, the R&D favourable deduction scheme, and patent boxes, are seldom-contested tax expenditures that are the biggest tax expenditures for many big economies.

• 2. Western legal systems follow the OECD and the World Bank and welcome tax incentives for research and development (R&D) as a sound innovation policy. In the European Union, nearly all Member States adopted tax subsidies for R&D expenditures. Belgium recently enforced and reformed its incentives for innovation, hereby following the Base Erosion and Profit Shifting initiative from the OESO.

INTRODUCTION

• 1. Recently, the availability and variety of innovative technologies, which support the provision of financial services, have increased considerably. Examples thereof range from biometric technology, virtual currencies, artificial intelligence and cloud computing to distributed ledger technology. During the past few years, we have not only been confronted with such terms, but also with the new buzzword ‘FinTech’. FinTech, the abbreviation of ‘Financial Technology’, is defined as ‘technologically enabled financial innovation that could result in new business models, applications, processes or products with an associated material effect on financial markets and institutions and the provision of financial services’. It is, in other words, an umbrella term for all the innovative products, services, processes, functions and distribution channels that have been created in the financial sector in order to respond to the needs of today's consumers. Financial technology – in the shape of algorithms, AI, ML and/or big data analytics – is frequently used when providing consumer credit. Creditors may use financial technology and automation to provide financial advice to the consumer asking for credit, thus guiding them in making a credit decision.

Before concluding a credit agreement, creditors are obligated to perform a thorough assessment of the creditworthiness of said consumer, based on sufficient information obtained from the consumer, where appropriate and on the basis of a consultation of a relevant database, where necessary. This assessment helps them decide whether this individual should be granted credit and allows them to verify whether the consumer will be able to meet his or her (financial) responsibilities under the credit agreement. This obligation to evaluate the financial and personal situation of each consumer who wants to acquire a loan, can be met through automated techniques, which require no or very little human intervention and instead rely on computer-based algorithms. Some of these automated techniques are called ‘credit scoring techniques’.

• 2 This chapter tries to answer the central question of whether consumer protection is fully guaranteed under the European Consumer Credit Directive, given the new digital reality and the extensive use of credit scoring techniques to assess creditworthiness.

INTRODUCTION

• 1. Artificial Intelligence has become a hot topic due to major advances in the field. However, many people participate in the debate without having the necessary understanding of the subject. In this chapter, we will explain some basic concepts of AI that may be useful for legal scholars and practitioners. It will provide readers with the necessary background to fully understand the impact of AI on law.

First, we provide a clear definition of AI and discuss the Turing Test. This test was a first controversial attempt to measure machine intelligence (part 2). We then focus on the working of AI. We consider two main AI approaches, namely knowledge-based and data-based learning. The latter is gaining importance every day, mainly due to the massive production of data by the Internet of Things (IoT). Machine learning (ML) can be considered the core of the data-based approach. One very popular ML method is the artificial neural network (ANN), which is described as well. We briefly discuss how it works and focus on its evolution into deep learning (DL). This evolution results from the increased data production and computing power. While DL has been a quantum leap for AI, it also has some drawbacks. These will be covered as well (part 3). AI has several sub-disciplines, many of which rely on ML. We briefly discuss search algorithms, computer vision, natural language processing (NLP), speech processing and agents (part 4). Having touched upon the foundations of AI, we subsequently focus on the wide range of areas and fields in which AI is already used. We discuss the current state of the art and expected evolutions in transportation, robotics, healthcare, education, public safety and security, art and entertainment, and law. We also look at the more distant future of AI (part 5). We conclude this chapter with some considerations regarding the ethical and safety aspects of AI (part 6).

DEFINING AND MEASURING AI

DEFINITION: WHAT IS AI?

• 2. Nowadays, AI is ubiquitous in the news. This results in a huge number of articles and (academic) papers on this technology. These texts are usually either overly optimistic or pessimistic1 regarding the possibilities as well as the dangers and challenges of AI.

INTRODUCTION

• 1. Artificial intelligence is becoming increasingly important in our daily professional and social lives. Although the use of AI systems has many benefits for a variety of sectors, different legal challenges remain. Some of these challenges are extensively discussed in this book. In this chapter, we will focus on the application of liability for damage caused by AI systems. The importance of liability and AI systems has already been mentioned in several recent documents issued by the European Union (EU). The White Paper on Artificial Intelligence, for instance, stresses that the main risks related to the use of AI concern the application of rules designed to protect fundamental rights as well as safety and liability-related issues. Scholars have also concluded that ‘[l]iability certainly represents one of the most relevant and recurring themes’ when it comes to AI systems.

• 2. This emphasis on liability is not surprising considering that AI systems will increasingly cause damage. Reference can be made to recent accidents involving autonomous vehicles. The autopilot of a Tesla car, for instance, was not able to distinguish a white tractor-trailer crossing the road from the bright sky above, leading to a fatal crash. A self-driving Uber car recently hit a pedestrian in Arizona. The woman later died in the hospital. A robot also attacked and injured a man at a tech fair in China. A surgical robot at a hospital in Philadelphia malfunctioned during a prostate surgery, thereby severely injuring the patient. In February 2015, a South Korean woman was sleeping on the floor when her robot vacuum ‘ate’ her hair forcing her to call for emergency help.

These examples show that accidents may happen despite optimising national and supranational safety rules for AI. This is when questions of liability become important. Nevertheless, the application of liability regimes for damage caused by AI systems can be challenging. The characteristics of AI systems such as opaqueness, autonomy, connectivity, data dependency or self-learning abilities make it difficult to trace back potentially problematic decisions made with the involvement of such systems.

INTRODUCTION

• 1. Arbitration is a method of alternative dispute resolution in which parties give up their right to have a legal dispute decided by a state court. Instead, they agree to authorise a so-called arbitral tribunal (composed of ‘private’ arbitrators appointed by the parties) to decide their dispute by rendering a binding decision. The decisions of arbitral tribunals are called arbitral awards and can be enforced if necessary.

• 2. One of the main selling points of arbitration is that it allows the parties to tailor the procedure to the specific needs of their dispute. More so than is the case for proceedings before state courts, the flexibility of arbitral procedure allows for the elimination of unnecessary procedural steps, saving both time and cost.

Compared to the technological revolution of the last decades, however, arbitrations are still ‘conducted in substantially the same way as they were 50 years ago’. Of course, standard applications such as e-mail and telephone or video conferences are already routinely used. Briefs are also usually created and submitted electronically, and increasingly include hyperlink references to the file or hearing transcript. Even slightly more advanced processes (such as identification, collection and transmission of responsive documents for the purpose of document production) are now conducted electronically and sometimes even generate new questions (e.g. in relation to the metadata accompanying such ‘electronically stored information’ or ‘ESI’). Generally speaking, however, this cannot be considered a true revolution.

• 3. As far as revolutionary technologies go, the recent buzz surrounding artificial intelligence has leftfew areas of the law unaffected. International arbitration is no exception.

Indeed, AI-based applications can already provide support in many different ways throughout the course of a legal dispute. They can, for instance, assist in the analysis or even in the conclusion of contracts; they can help the parties in making strategic decisions based on data-analytics (such as which arbitrator to appoint); they can help counsel in the analysis and the drafting of submissions; and speedily process large amounts of data in electronic discovery.

INTRODUCTION

• 1. AI, robotics and other forms of smart automation have the potential to bring great economic benefits – up to $15 trillion to global GDP by 2030 – causing a major shiftin the global economy. It is seen as a key driver and component of the Fourth Industrial Revolution. This Revolution is more transforming than any other industrial revolution we already experienced so far. It challenges our ideas about what it means to be ‘human’. The influence of AI can already be seen on the labour market (e.g. robotisation of work) as well as in public (e.g. facial recognition) and private spaces (e.g. virtual assistants at home). The integration of AI within our daily routines makes it hard to imagine life without it.

  2. Many governments are also investing in AI. China is already a global leader in AI and will probably achieve its aspired $150 billion AI investment by 2030. The United States (US) acknowledged that AI will be their second highest R&D priority after the security of the American people, investing up to $2 billion between 2018–2023 for the advancement of AI. In Europe, primarily France and the United Kingdom (UK) are at the AI forefront, with France investing $1.5 billion and the UK $1.3 billion in AI-related research. As a consequence, the geopolitical implications of AI should not be underestimated. This is even more so considering the relationship between many AI applications and automated warfare, providing any AI-investing country with a potential military advantage. The increased investments in AI research and development show the importance of AI and its potential ability to rebalance world power.

  3. Considering that AI systems affect many aspects of our lives, it is necessary to question what the role of human rights will be in the AI and robotics era. To assess this role, a helicopter overview of the way in which human rights are or can be impacted by the commercialisation and use of AI is provided (part 2). Several issues addressed in this chapter are covered more thoroughly in other parts of this book..

INTRODUCTION

• 1. Business entities currently employ AI and other algorithmic techniques in essentially all sectors of the economy in order to influence potential customers. The concept of AI is discussed elsewhere in this book. This contribution is more concerned with what is happening on the market under the label ‘AI’ and how this may affect those who are generally labelled as consumers. After all, the focus of legal research is not so much on ‘new’ technology itself, but rather on the aspects of social life that this technology makes newly salient.

To that end, I will first identify and categorise some of the ways in which business entities employ what is commonly referred to as AI as well as the risks and benefits of such uses (part 2). For this, I will rely on the findings of the ARTificial intelligence SYstems and consumer law & policy project (ARTSY Project) conducted by the European University Institute in Florence under the supervision of professor Hans Micklitz. Subsequently, I will examine how the legislator intends to adapt consumer policy to the changing circumstances created by these previously mentioned developments (part 3). I will limit this study to European Union consumer policy as the Belgian legislator is likely to adopt this approach. I will then examine some of the hurdles (AI-driven) autonomous agents present to consumer autonomy as well as the question to what extent and how this can be dealt with within the current consumer law framework (part 4). In particular, I will discuss a number of market practices which are closely related to the advent of autonomous agents. In this regard, I will rely on the key issues in the consumer domain as defined in a briefing document to the European Parliament prepared by one of the researchers of the ARTSY Project. I will not elaborate on consumer privacy as privacy considerations are discussed elsewhere in this book. Finally, I will recapitulate my findings and contemplate on the nature of consumer rights in the era of AI (part 5).

BENEFITS AND RISKS OF AI AS A MARKET TOOL

• 2. A sectoral analysis prepared within the framework of the ARTSY Project shows that the use of AI is booming in several domains.