-
PDF
- Split View
-
Views
-
Cite
Cite
Lutz-Christian Wolff, Artificial Intelligence ante portas: The End of Comparative Law?, The Chinese Journal of Comparative Law, Volume 7, Issue 3, December 2019, Pages 484–504, https://doi.org/10.1093/cjcl/cxz020
Close - Share Icon Share
Abstract
Artificial intelligence (AI) can do many things that were not thought of some years ago and that are unimaginable for non-AI experts even today. In contrast, it is relatively easy to understand that AI can be used to compare contents and structures of laws and legal documents. In fact, the comparative abilities of AI are the reason why AI is now playing an increasing role—for example, in due diligence exercises where contracts, documents as well as other materials and legal data of target companies are benchmarked against standard patterns. If the ‘ability to compare’ is one of the core features of AI it is only natural to assume that AI is an ideal tool to conduct comparative law work. This article explores if this assumption is correct. This article first highlights key features of the comparative law work process, which, for some strange reason, is hardly ever discussed in the legal literature. This article describes and analyses the different stages and investigates which parts can (or cannot) be conducted by AI. It also asks if AI will—within the scope of its comparative abilities—in fact, ‘take over’ from human comparatists. On the basis of the findings, this article concludes that it is more likely than not that comparative law work will, in the future, be AI based.
Introduction
Artificial intelligence (AI) can do many things that were not thought of some years ago and that are unimaginable for non-AI experts even today. In contrast, it is relatively easy to understand that computer and cloud-based systems, including AI, can be used to compare contents and structures of laws and legal documents. These ‘comparative abilities’ of AI are, for example, the reason why AI is now playing an increasing role in due diligence exercises where contracts, documents, and other materials of target companies are searched and benchmarked against keywords and standard patterns.1 Because the ability to compare appears to be one of the core features of AI, it is only natural to assume that AI is an ideal tool to conduct comparative law work. This article explores if this assumption is correct.
It is common understanding that AI ‘will increasingly replace repetitive jobs. Not just blue-collar work, but a lot of white collar work…40% of jobs in the world, including truck drivers’, telesales people’s, security guards’ and even radiologists’ will become displaceable by technology over the next 15 years.’2 Ultimately this article, therefore, also addresses the question of whether AI will make legal comparatists redundant.
While it has been claimed that ‘[t]he limits on AI’s use are often more about the imagination of the users, than the technology itself’,3 it must be acknowledged that ‘we still live in a world without successful operating AI in law systems’.4 In particular, at present, AI is not (yet) comprehensively used to conduct comparative law studies. In fact, if this was the case, there would be no need to explore related options in this article.
AI is an evolving concept, and the technical abilities and limits of AI5 are widely discussed amongst information technology experts.6 This article refers to related aspects, where appropriate, on the basis of a literature review and discussions with computer scientists. However, this article does not aim to join the debate about the current or future technical standard of AI. It rather attempts to conceptualize the options AI offers (or may offer) from the viewpoint of the comparative law work process. For example, on the one hand, it has often been suggested that comparative law studies or, at least, certain parts of the comparative law work process are driven by the personal interest of involved researchers.7 On the other hand, it is commonly, but not necessarily correctly,8 assumed that AI does not have emotions and interests.9 Therefore, if the personal interests of researchers are indeed so important for comparative law work, then it seems that AI is conceptually precluded from related tasks.
The structure of this article is as follows: In its first section it briefly recalls what AI is. It then highlights core stages of the comparative law work process. For each of these stages this article explores whether related work can potentially also be done by AI. It also asks if AI will, within the scope of its comparative abilities, in fact, ‘take over’ from human comparatists. This article concludes with final remarks of a more general nature.
Various aspects of the use of AI in relation to particular legal fields and work types have been discussed in the past with different degrees of depth.10 However, the role of AI in the context of comparative law has never been addressed. This topic, however, is highly important not only because of the increasing significance of AI in the modern world in general but also because it offers a completely new perspective of the work process of legal comparatists and thus allows for a reconsideration and refinement of comparative law methodology.11 Last but not least, a lot of what is discussed in this article may sound like science fiction to the comparative law world. It is, however, plain vanilla talk for computer scientists and, therefore, deserves all the more attention.
Artificial intelligence
‘Artificial intelligence’ is not a technical term, and it is, therefore, not used in a standardized way around the globe. In fact, since the term was first introduced in the 1950s, its definition has seen several changes and developments.12 While there is a growing literature attempting to delineate the notion of AI from various angles, ‘[t]oday, modern dictionary definitions focus on AI being a sub-field of computer science and how machines can imitate human intelligence (being human-like rather than becoming human)’.13 Accordingly, for the purpose of this article, ‘AI’ shall refer to machines and machine systems—normally, computer systems that mimic human intelligence. Commonly, a distinction is drawn between ‘weak AI’ (also ‘narrow AI’), ‘strong AI’ (also ‘general AI’), and ‘super AI’.
Weak AI
is the AI that exists in our world today…[Weak] AI is AI that is programmed to perform single tasks – whether it’s checking the weather, being able to play chess, or analyzing raw data to write journalistic reports…within a pre-determined, pre-defined range.14
In contrast, strong AI ‘can successfully perform any intellectual task that a human being can’.15 Super AI ‘will surpass human intelligence in all aspects. This is the type of AI that many people are worried about’16 and that will be discussed in the last section of this article.17 Often, AI and ‘automation’ are used interchangeably.18 For this article, however, it appears important to distinguish both concepts:
In general usage, automation can be defined as a technology concerned with performing a process by means of programmed commands combined with automatic feedback control to ensure proper execution of the instructions. The resulting system is capable of operating without human intervention.19
Automation
[Automation] has a single purpose: to let machines perform repetitive, monotonous tasks. This frees up time for fleshly humans to focus on more important tasks that require the personal touch. The end result is a more efficient, cost-effective business and a more productive workforce…. Essentially, it’s a machine that’s smart enough to follow orders.20
The difference between AI and automation has been described as follows:
Automation is basically making a hardware or software that is capable of doing things automatically. …On the other hand Artificial Intelligence is a science and engineering of making intelligent machines…AI is all about try to make machines or software mimic human behavior and intelligence. …Automation can or can not be based on Artificial Intelligence.21
It is a very important feature of AI that it imitates the human learning process.22 AI grows its capabilities like humans who, from early childhood, adopt very basic knowledge and skills and develop them, over time, to advanced levels.23 During the increasingly sophisticated learning processes, which themselves are constantly refined, anything new triggers new learning activities until the new feature is understood and can be applied and also be independently developed further. The more new that things AI encounters, the easier and faster is the learning process. The ability of software to learn and train itself is normally referred to as ‘machine learning’ (ML):24
Internet AI, which started in 1998, enables Amazon to predict what its users might want to buy and powers Facebook to suggest what its users might want to follow. In Business AI starting in 2004, data have become goldmine for banks, insurance companies and hospitals to improve decision-making and maximize productivity. Perception AI started in 2011 when we gave AI the ability to see, hear and speak. In the fourth wave of Autonomous AI starting in 2015, AI is given the ability to respond to the world around it, to move intuitively, and to manipulate objects as easily as a human can.25
AI is sometimes broken down into particular AI technologies. Examples are ML and natural language processing (NLP). NLP stands for software systems that can read ‘natural language’ embedded in a text.26
Can AI conduct comparative law work?
General
The previous section introduced basic features of AI. This section now explores to what extent AI can be used to conduct comparative law work. Comparative law, of course, does mean ‘different things for different people, and each of these meanings can, in and of itself, be scientifically acceptable’.27 The following discussion shall not be limited to any specific understanding of comparative law but, rather, adopts the broadest scope possible—that is, it regards comparative law as the legal discipline that engages in any kind of ‘comparison of the different legal systems of the world’.28 Moreover, comparative legal research can relate to legal systems in general, special areas of law, or particular rules, concepts, approaches, institutions, and legal cultures.29 Again, this article is not limited to any of these features but shall be ‘all-inclusive’. In order to allow for a focused discussion, however, it only refers to ‘legal systems’ unless otherwise stipulated.30
At its core, the comparative work process comprises four stages—namely, the setting of methodological parameters that guide the comparative study; the collection of data and other information of the legal systems to be compared; the act of comparing as such; and, finally, the analysis of the findings. In the following, these four stages are introduced briefly and then analysed respectively in respect of the question of whether related work can, potentially, also be conducted by AI.
Setting methodological parameters
Methodological questions normally do not receive much attention when comparative law work is conducted.31 In contrast, comparative law work should always start with the establishment of a methodological framework within which the comparison is conducted and that thus determines the research direction. Comparative work that lacks such a methodological framework can of course still generate meaningful output. However, such output would be the result of random work and therefore not be methodologically sound.32
In order to establish a methodological framework for a comparative law project, five core questions need to be answered. First, what is the goal of the comparison? Second, what shall be compared? Third, which jurisdictions should be selected for the comparative purposes? Fourth, how can comparability be ensured? And, finally, how should the comparative work be conducted? In the following, the significance of each of these core questions is recalled one by one. Subsequently, it shall considered if they can also be answered by AI—that is, if AI can set the methodological parameters that guide comparative law work.
Like research work in any other area, comparative law research should pursue particular goals in order to avoid randomness. These goals should be reflected in the specific research questions that the study aims to answer.33 In contrast, comparative law work without goals may be of aesthetic value or entertain the researcher’s curiosity, but would be random and thus methodologically flawed.34
Comparative law research can pursue very different goals and it is the prevailing opinion that an exclusive enumeration may not even be possible.35 However, four goals stand out. First, many comparative law projects aim to improve one or more legal systems—that is, they are reform-oriented.36 Second, comparative law studies are often done (solely) to achieve a better understanding of the involved legal systems—that is, they are undertaken for the sake of knowledge advancement.37 Third, comparative work is often necessary in the day-to-day professional legal work—for example, for lawyers to advise clients on cross-border legal issues as part of the private international law process or for courts,38 and for legislators, when considering foreign law, to find solutions for domestic legal problems.39 Fourth, comparative law work is of course also important in the context of legal education and can thus be aimed at pedagogical goals—for example, when explaining particular legal systems in educational contexts by benchmarking them against those applied elsewhere.40
Comparative law research can entail micro or macro comparisons.41 In other words, it can relate to legal systems as such, special areas of law or particular rules, their sources and their application, concepts, approaches, institutions, and to legal cultures as well as the law’s contextual setting.42 The decision of what to compare forms part of the methodological setting of, and is to be determined in line with the goal(s) of any comparative law project.43 Furthermore, because the goals of comparative law are indefinite,44 the selection of the objects of comparative work is necessarily also open-ended. For example, if comparative law work aims at legal reform then the precise reform goals will determine the object of the comparative study—for example, jurisdictions as such, a particular area of law, some kind of legal institution, or the legal culture. The same is true for professional legal work that employs comparative law tools. Here, the question of what to compare is determined by the very outcome that must be achieved. Moreover, if comparative law work is conducted for pedagogical reasons then the question of what to compare will have to be answered closely in line with the aspired educational outcomes.
After having determined the object of a comparative law study, it must be decided which jurisdictions to consider for the comparative purposes.45 Comparative law work can be horizontal, vertical, or third dimensional. It is horizontal when legal systems that are located at the same level, such as the legal systems of different countries or of different states of the same country are compared. Comparative law research takes a vertical approach when legal systems located at different levels are compared—that is, the comparison of national systems with international systems46 or the comparison of state systems with federal systems.47 Time can be seen as the third dimension of comparative law. Third-dimensional comparison is one form of legal history research and involves legal systems at different points of time or time periods.48 Whether horizontal, vertical, or historical, it must be decided which jurisdictions legal systems are to be chosen from for the comparative purposes pursued.49
While such selection is very important from a methodological point of view, it is interesting to note that many comparative law studies do not explain or even justify their choice.50 In contrast, if comparative law work should pursue particular goals,51 then the jurisdictions to be compared in relation to particular legal systems must be selected accordingly—that is, the selection must support these goals.52 And, in order to ensure transparency and to allow for a verification of the project’s methodological viability, the selection must be justifiable and should be explained.53 In fact, it is rather obvious that the choice of suitable jurisdictions is crucial in order to meet the goals of a comparative law project—that is, for its success.54 But how can jurisdictions be identified in pursuance of particular goals?55
In practice, very often comparatists’ personal features, such as language ability, familiarity with a particular legal system or tradition, or access to primary or secondary sources,56 determine the selection of jurisdictions for comparative purposes. If personal features of this kind are, as a matter of principle, unrelated to the goals of the study, such a selection would be random and not methodologically sound.57 The same is true if time and cost constraints determine the choice of jurisdictions to be compared. Consequential limitations require acknowledgement and justification.58
Comparative law research has, for a very long time, relied on the concept of legal families (also legal traditions)59—that is, the focus was on jurisdictions that are representative60 so as to rationalize the comparative law work. It is, however, highly disputed which legal families can be acknowledged and why.61 Furthermore, in times of globalization, pure legal systems hardly exist.62 Therefore, the legal family concept allows for interesting insights but cannot offer conclusive answers to the question of which jurisdictions are to be selected for comparative purposes.63
Some commentators have suggested that jurisdictions with very similar or very different systems,64 or systems that are at the same stage of development, should be selected for comparative purposes. However, the degree of difference between legal systems or the stage of development65 does not justify the selection of related jurisdictions for comparative purposes if they are not in line with the project’s research goals.
In contrast, if comparative law work aims at reform, state-of-the-art systems may be selected to serve as models.66 If legal reform is pursued through harmonization or unification,67 then the underlying comparative analysis should cover all the concerned jurisdictions, and probably other jurisdictions, for benchmarking purposes.68 As far as comparative work is part of the day-to-day judicial process the choice of the jurisdictions will be mandated by the objectives of the involved legal work. Similarly, if comparative work is done for pedagogical reasons the learning outcomes to be achieved should drive the selection of the jurisdictions.
Comparatists also have to consider the problem of comparability—that is, the question of which legal system in one jurisdiction corresponds to which in another.69 While this is commonly acknowledged, it is debatable how comparability can be ensured.70 As far as macro-comparison is concerned—that is, when jurisdictions as such are the object of the comparative endeavours—reference can be made to the discussion in the previous paragraphs. For micro-comparison, traditional comparative law defines the goals of micro-comparison on the basis of ‘functional equivalence’—that is, comparability implies that the compared rules, institutions, and principles must have the same role71 or be ‘solving the same real life, socio-economic problem’.72
Finally, legal systems can be compared in different ways with different starting points and different modes how to move forward.73 For the methodological set-up of a comparative law project it must, therefore, be decided which of these shall be pursued. Traditional comparative studies often engage in doctrinal comparison—that is, they seek to provide a systematic comparative analysis of the rules and underlying principles governing a particular legal system.74
Doctrinal comparison can be supplemented by a contextual comparison or the contextual comparison of legal systems can be conducted independently. Contextual comparison is based on the notion of path dependence of legal rules—that is, the understanding that law is not free-standing but acts within and relies on a particular political, economic, cultural, and legal context.75 Contextual comparison thus means the analysis of the status and the functioning of legal systems within particular settings.76 Comparative law work can also be empirical in nature: ‘Empirical research develops findings through the collection of data gathered through observation that is usually used to test research hypotheses.’77
Moreover, comparative law work can relate to legal systems in different time periods. This kind of comparative legal history work, however, would also have to rely either on doctrinal or on empirical analyses and has itself been regarded as a contextual method.78 From this point of view, it may be questionable if comparative legal history research qualifies as an independent methodological category.
The decision of how to compare, again, should not be random. In contrast, different goals and purposes of comparative law work demand particular approaches or a combination of them. In other words, the selection of a particular approach requires justification and the limitation to one or more approaches must be explained, and resulting restrictions in terms of the scope of the study must be acknowledged for the sake of transparency.
The previous sections have summarized core questions that must be addressed when establishing a methodological framework to guide comparative law work. In the following it is now discussed if and which of these core questions can be answered by AI. First, it appears that AI would not be able to set the goals of comparative law work. The setting of comparative law goals starts with a research idea on the basis of a general understanding of the concerned legal systems. It thus requires independent initiative and creativity. Weak AI,79 as it exists for the time being, is not able to enter into the comparative law work process without the instruction to do so. In other words, AI lacks the ability of self-motivation: ‘The creative and the complex, such as those done by scientists and artists, economists and CEOs, will be protected, because “AI can optimize but not create.”’80 One may, of course, speculate if this will change in the future as ‘computers can, in theory, emulate human intelligence, and exceed it.’81 This will be discussed below in the last section of this article.82
Second, as explained above, the goals of comparative law work determine the precise object of the comparison,83 the choice of the jurisdictions from which the object of the comparative law work is to be selected,84 and also the way in which the actual comparison is to be carried out.85 In other words, once the goal(s) of a comparative law project are determined, these three methodological parameters should follow accordingly—that is, they should be set in line with what is most suitable to reach the stated goal(s). It, therefore, appears that decisions in these three domains do not require any independent cognitive input, but are rather a natural consequence of the definition of the goal(s).86 Leaving aside the technical complexity of AI systems that are able to determine objects of, jurisdictions relevant for, and the most suitable way to conduct comparative work, conceptually AI should be in a position to set these parameters.
Finally, the issue of comparability needs to be decided on the basis of an understanding of the concerned legal systems, and there is no reason why AI should be less able to accomplish this task than human comparatists if fed with relevant data and other information. In fact, due to its ability to process large amounts of data and information at high speed, AI should be much faster and more efficient than any human endeavour can ever be in this regard.
Collecting data and other information
The comparison of legal systems is based on related data and other information.87 Data and information of this kind will often be readily available for retrieval from sources like libraries and online databases. Alternatively, if this is not the case, data and information must be collected or else a comparison cannot take place.88 The retrieval or collection of data and information related to legal systems to be compared has, again, to be in line with the goals of the comparative law study and forms an important part of the comparative law work process. The computer-based collection of data and other information is, nowadays, standard in many areas.89 AI, therefore, should also have the ability to collect information and data in the context of the comparative law progress.90 However, three aspects deserve to be highlighted.
First, even simple search engines can collect laws, rules, and legal documents. AI can go further and analyse these laws, rules, and legal documents comprehensively and identify special features. Furthermore, AI is able to identify not only the rules set out in statutory law and contract documents but also those developed by case law.91 AI can even consider application practices in relation to particular legal rules on the basis of available data and information as a result of its constant learning and self-training process.92 The more access AI has to data and other information, the quicker it will be able to teach itself how to improve the collection process and thus bring itself to perfection: ‘What drives both automated systems and AI is the same thing that drives businesses: data. Automated machines collate data; AI systems “understand” it.’93
Second, it appears that AI would not even require specific instructions to retrieve or collect data and other information related to the legal systems to be compared. In contrast, the setting of the methodological parameters of a comparative study should enable AI to determine which information and data are required to reach the stated goals of the comparative study and then to retrieve or collect them.
Finally, for the time being, AI does not seem to be able to engage in information collection that is traditionally done through person-to-person interaction, such as interviews or behavioural observation. One may speculate, however, whether its undisputed data mining ability will also enable AI to deduct related information from other sources to generate knowledge that is at least at the level of that produced by human comparatists.
Comparing
The actual comparison stands at the core of comparative law work.94 According to the Cambridge Dictionary, comparing means ‘to examine or look for the difference between two or more things’.95 In fact, since comparing can also entail the search for similarities, comparative law work should be seen as the examination or the search for differences and similarities between two or more legal systems.96 How such comparisons can be conducted was highlighted in a previous section. Once the methodological parameters of a comparative law study are set, the actual comparison does not require any additional cognitive input. As mentioned at the outset, it therefore stands to reason that AI can compare information and data related to legal systems.97
Drawing conclusions
Comparative law work, of course, is not limited to the identification of differences and similarities between legal systems. In contrast, in line with their stated goals, comparative law projects will normally go further and engage in an analysis of the findings.98 Often this analysis will then inspire recommendations for changes of one or more legal systems.99 In other words, a comparative law study will not just aim to consider if X is different from or similar to Z but will also explore the reasons for the differences, the similarities, and the actual consequences as well as often also assess whether option X or Z are superior in given circumstances100 and if this leads to the requirement of changes.
The currently existing limits of AI’s ability to engage in legal analysis were mentioned above and have been widely discussed by commentators in the past.101 For the question of whether this means that AI is generally unable to draw conclusions that go beyond the (simple) identification of differences and similarities of legal systems, it must first be considered that AI has developed at an almost dramatic pace in the last 10 years. Comments on what AI is able to achieve made even in very recent times may, therefore, no longer be valid.102 More importantly, conclusions to be drawn by AI on the basis of comparative findings would not at all be free-standing. They would rather have to be in line with the methodological parameters set for the study. Because the methodological parameters determine if and how conclusions are to be drawn in order to achieve the stated goal(s), AI should be able to analyse differences and similarities and draw conclusions independently. And this includes, for example, the identification of reform needs and the development of suggestions on how these needs can be addressed, the translation of comparative findings into a certain teaching pedagogy, as well as solutions regarding particular practical problem by way of considering solutions offered in other jurisdictions.
Deducing value judgments and considering path dependence
The above sections discussed AI’s abilities during the different stages of the comparative work process. Two additional aspects of a more general nature require attention in this regard.103 First, the design of law is not only a result of objective factors but also represents particular value systems, thus reflecting underlying jurisprudential, cultural, political, and societal concepts.104 Legal comparatists must take these concepts into account at any stage of the comparative law process, and it must be decided if AI would be able to do the same. In this regard, it is also important that many comparative law issues are still unsettled,105 and comparatists have to take positions, often on the basis of their own value judgments. Again, it must be asked if AI is able to replicate the decision-making process.
Second, law functions differently within different political, economic, social, cultural, and legal environments.106 This so-called path dependence of the law must be considered during the comparative law work process.107 Vice versa, the operation of the law within a given environment—that is, the existing legal culture—must be taken into account as well. The question, again, is if AI is able to do the same.
For the question of whether AI can deduce value judgments underpinning particular legal systems, it must be considered that any data and other information regarding such legal system will be an expression of these value judgments. Therefore, comprehensive access to these data and other information should enable AI to deduce the underlying value judgments and use them for comparative law purposes. The same is true in regard to the environment within which law operates—that is, related data and information should allow AI to understand the law’s contextual settings as well as how it operates within the given environment. Even if one was to question whether data and other information will allow AI to generate a complete picture, it must be remembered that AI would only replicate what human comparatists do. Human comparatists also have to rely on available data and other information in order to deduce the values underpinning legal systems and to understand their contextual settings.
The ability to deduct underlying value judgments and contextual features, of course, does depend on comprehensive access to information and data. Therefore, if ‘comprehensive access’ is understood to mean the possibility to retrieve and process 100 per cent of the data and information related to all legal systems of the world, it must be concluded that AI is not there (yet). However, as explained above,108 AI has advanced data collection and processing abilities. It is safe to assume that these abilities do, or at least will soon match the abilities of human comparatists. In fact, AI should be in a position to collect and process data and information in a much faster, more automated, and thus less error-prone way as compared with human comparatists.109
Finally, understanding value systems and contextual features and processing them for comparative law purposes implies that AI relies on the status quo. In contrast, AI experts and other scientists have anticipated with great concern the possibility of super AI,110 which has the ability to go beyond the status quo—that is, that AI may at one point in the future be in a position to change existing or create new value systems and contextual features on its own. This possibility is discussed in the next section.
Will AI Take over Comparative Law Work?
General
It was the conclusion of the previous sections that AI is not (yet) able to independently define the goals of, and thus to initiate, comparative law work. However, it must be assumed that AI can otherwise perform any of the involved tasks at least at the performance level of human comparatists. This section now discusses what this means for the future role of AI in the context of comparative law work.
While it is not the goal of this article to assess if technical advancement will allow AI to overcome currently existing obstacles in the future,111 it is noteworthy that AI experts have estimated that a ‘perfect AI world’ is not at all science fiction but could become reality in as early as 10 to 30 years.112 The following discussion will assume such a ‘perfect AI world’—that is, the situation when AI is not subject to any restrictions—and discuss what this means for comparative law as a legal discipline.
Comparative law in a perfect AI world
Currently, the lack of comprehensive access to data and other information is one of the main obstacles for AI to enter into full swing.113 In contrast, if AI had such comprehensive access to data and information of all legal systems of the world, then AI could engage in a permanent process of data and information retrieval, comparative analyses of these data and information, the identification of societal needs,114 and the development of solutions to address these needs or problems through the comparative law approach.115 In a perfect AI world of this kind, AI may even be able to implement its own solutions.116 In other words, AI could act and react completely independently—within the comparative law and other contexts—thus forming an AI-based smart legal system that is in a constant process of reforming itself.
In a perfect AI world, even doubts regarding AI’s ability to set goals of comparative law work independently and thus to initiate the comparative law process could be dispelled.117 As indicated in the previous paragraph, in a perfect AI world, reform goals would not have to be set; they would rather be system inherent. In other words, an AI-based smart legal system would constantly identify and address the problems of legal systems. As far as practical questions arising in the judicial process or pedagogical issues118 are concerned, AI would be able to initiate the comparative law process as soon as access to related data and information indicate the need for a solution.
Interestingly, the situation appears to be different as far as (pure) knowledge advancement is concerned. As discussed above,119 knowledge advancement is often seen as one of the main drivers of comparative law projects. But knowledge advancement can hardly be seen as a goal of AI-driven comparative law work. First, it is difficult to imagine particular interests pursued by AI. AI does not have a will, does not have emotions, and does not have interests. Second, and more importantly, in a perfect AI world, full knowledge regarding all of the world’s legal systems would be available in the form of data and information and could be processed in real time. In other words, comparative knowledge would not have to be developed; it would just have to be retrieved. And, indeed, this may imply the end of comparative law as a legal discipline.
Preventing AI from taking over comparative law work?
The (potential) ability of AI to operate independently from human beings—that is, the creation of super AI120—has led to great concerns and is one of the core topics of the ongoing general AI debate. The late Stephen Hawking famously summarized the worries as follows:
Success in creating effective AI, could be the biggest event in the history of our civilization. Or the worst. We just don’t know. …Unless we learn how to prepare for, and avoid, the potential risks, AI could be the worst event in the history of our civilization. It brings dangers, like powerful autonomous weapons, or new ways for the few to oppress the many.121
While it is beyond the scope of this article to discuss the risks of AI as such, it must be considered whether related concerns are valid also from the comparative law point of view. Why should there be anything wrong with an AI system that conducts comparative law work independently, with the utmost efficiency, for the benefit of potentially perfect solutions? In fact, as long as AI functions within the boundaries defined by the existing legal system(s)—that is, as long as AI remains neutral—AI appears to be nothing but a perfect tool to conduct comparative law research.
The situation, however, would be completely different if AI failed to stay neutral. ‘Malicious AI’ of this kind could be created by AI architects intentionally or by error,122 or ‘good AI’ could simply run out of control.123 And this, of course, is also possible in the context of comparative law work. For example, AI could decide that economic development always has absolute priority, that human rights and the rule of law are not to be considered any longer, and that the goals of comparative law and the comparative law work process have to be adjusted accordingly. In fact, the risk that AI moves outside existing value systems is rather manifest because cognitive independence is the very idea of AI.124
From the viewpoint of comparative law, one might of course consider preventing related problems by simply continuing comparative work in the traditional way—that is, to ignore all of the options AI may offer. However, this approach does not make much sense in light of the fact that comparative law work conducted by AI will be faster and more precise and thus more efficient than what can ever be achieved by human comparatists.
AI experts are sometimes trying to disperse AI-related concerns by pointing to the possibilities to control AI: ‘After all, AI is merely a tool that we, as humans, control and run. We are the ones who have free will and is up to us to direct the plots of the AI story.’125 But the comments by Stephen Hawking and others show that even experts in the field are not convinced that AI can ultimately be hindered from taking over.126
Conclusion and final remarks
This article has attempted to explore the future role of AI for comparative law work from a conceptual point of view. The analysis has allowed at the same time for the opportunity to (re)-consider the comparative law work process as a whole, a topic that finds surprisingly little attention in the legal literature.
This article has demonstrated that AI does not yet seem to be able to determine the goals of, and thus initiate, comparative law work independently. However, once the goals of a comparative law project are set, the subsequent stages of the comparative law work process—that is, the setting of other methodological parameters, the collection of data and other information to inform the comparative law work, the comparison as such, and the drawing of conclusions—do not require independent cognitive input but are guided by the pursued goal(s). Leaving aside the question of whether and when an AI system will be created to follow the whole comparative law work process through, AI, from a conceptual point of view, is already able to complete all of the tasks related to these subsequent stages.
Whether AI will be used for comparative law work in the future and what degree of independence AI will gain in this regard depends on many factors—last, but not least, on the willingness and ability of comparatists to let this happen. In this regard, it is important that AI experts assume that a ‘perfect AI world’—that is, a world in which technical restrictions of AI do no longer exist—is only 10 to 30 years away.127 Realistically speaking, it is therefore more likely than not that major parts of comparative law work will soon be AI based.
Many thanks go to Mahdi H Miraz for valuable comments regarding the computer engineering aspects of this article and to Jenny Chan for great research support.
Footnotes
See Anonymous, ‘Diligence Disrupted: Law Firms Climb Aboard the AI Wagon’ The Economist (12 July 2018) <https://www.economist.com/business/2018/07/12/law-firms-climb-aboard-the-ai-wagon> accessed 15 December 2019; Richard Tromans, ‘Legal AI: A beginner’s Guide’ 5 <https://blogs.thomsonreuters.com/legal-uk/2017/02/20/legal-ai-beginners-guide/> accessed 15 December 2015.
Lee Kai-fu, former Apple, Microsoft and Google executive and now chairman of Sinovation Ventures, quoted by Christine N, ‘Hear the Oracle Speak: Lee Kai-fu Forecasts the Future of AI’ Chinese University of Hong Kong Newsletter (19 April 2019) 4 <http://www.iso.cuhk.edu.hk/images/publication/newsletter/536/html5/4/#zoom=z> accessed 15 December 2019; Victor M Palace, ‘What If Artificial Intelligence Wrote This: Artificial Intelligence and Copyright Law’ (2019) 71 Florida L Rev 217, 240–1; Robert C Denicola, ‘Ex Machina: Copyright Protection for Computer Generated Works’ (2016) 69 Rutgers UL Rev 251, 255–6.
Tromans (n 1) 3.
Abdul Paliwala, ‘Rediscovering Artificial Intelligence and Law: An Inadequate Jurisprudence’ (2016) 30(3) Intl Rev L, Computers & Technology 107, 112; for the problems ‘that have come to light’, see James Franklin, ‘Discussion Paper: How Much of Commonsense and Legal Reasoning is Formalizable? A Review of Conceptual Obstacles’ (2012) 11 Law, Probability & Risk 225.
Palace (n 2) 222.
Cf J Storrs Hall, Beyond AI: Creating the Conscience of the Machine (Prometheus Books 2007) 250, estimating that perfect artificial intelligence (AI) systems may be available in the 2020s or 2030s.
Cf Esin Örücü, ‘Methodology of Comparative Law’ in Jan M Smits (ed), Elgar Encyclopedia of Comparative Law (Edward Elgar 2006) 443; Konrad Zweigert and Hein Kötz, An Introduction to Comparative Law, translated by Tony Weir (3rd edn, Clarendon Press 1998) 37; David J Gerber, ‘System Dynamics: Toward a Language of Comparative Law?’ (1998) 46(4) Am J Comp Law 719, 732; Nils Jansen, ‘Comparative Law and Comparative Knowledge’ in Mathias Reimann and Reinhard Zimmermann (eds), The Oxford Handbook of Comparative Law (OUP 2006) 305, 314–15; see also the more critical assessment in this article under the heading ‘Setting Methodological Parameters’.
Compare later in this article under the heading ‘Deducing Value Judgements and Considering Path Dependence’.
Lee Kai-fu (n 2) 4; Herve-Victor Fomen, ‘Can Artificial Intelligence Become Emotional?’ Insights (29 November 2018) <https://www.leyton.com/blog/?p=3193-can-artificial-intelligence-become-emotional> accessed 15 December 2019.
See EC Lashbrooke Jr, ‘Legal Reasoning and Artificial Intelligence’ (1998) 34 Loyola L Rev 287; Sean Semmler and Zeeve Rose, ‘Artificial Intelligence: Application Today and Implications Tomorrow’ (2017–18) 16 Duke L & Technology Rev 85; Ed Walters, ‘Read/Write: Artificial Intelligence Libraries’ (2017) 22 Am Assoc L Libraries Spectrum 21, 22: ‘Libraries also can use AI tools to help change the way that law school clinics deliver legal services’; Sarah J Moore, ‘Artificial Intelligence in the Workplace’ (2017) 31 Ohio Lawyer 18, 18; Cristian-Vlad Oancea, ‘Artificial Intelligence Role in Cybersecurity Infrastructures’ (2015) 4 Intl J Information Security & Cybercrime 59; Steven K Rainey, Brad Brown and David B Kirk, ‘Bots, Natural Language Processing, and Machine Learning’ (2017) 69 Tax Executive 39, 40: ‘Tax departments are beginning to understand and apply the great potential of intelligent automation to their core functions’; Benjamin Alarie, Anthony Niblett and Albert H Yoon, ‘How Artificial Intelligence Will Affect the Practice of Law, Artificial Intelligence, Technology, and the Law’ (2018) 68(1) U Toronto LJ 106, 108, citing John O McGinnis and Russell G Pearce, ‘The Great Disruption: How Machine Intelligence Will Transform the Role of Lawyers in the Delivery of Legal Services’ (2014) 82 Fordham L Rev 3041; Sherry Xin Chen and Mary Ann Neary, ‘Artificial Intelligence Legal Research and Law Librarians’ (2017) 21 Am Assoc L Libraries Spectrum16; Daniel B Evans, ‘Artificial Intelligence and Document Assembly’ (1990) 16 L Practice Mgmt 18.
For the lack of recognition of the importance of comparative law methodology, cf Vernon Valentine Palmer, ‘From Lerotholi to Lando: Some Examples of Comparative Law Methodology’ (2005) 53 Am J Comp Law 261, 262: ‘Some of the most widely read books on comparative law have virtually nothing to say about methodology and, perhaps in consequence, the rank and file may be described as naïve and unaware of methodological questions and issues’; for the fact that there is no commonly accepted definition of comparative law methodology, see J Paul Lomio, Henrik Spang-Hanssen and Henrik Wilson, Legal Research Methods in a Modern World: A Coursebook (3rd edn, Djøf Publishing 2011) 56.
Bernard Marr, ‘The Key Definitions of Artificial Intelligence (AI) That Explain Its Importance’ Forbes (14 February 2018) <https://www.forbes.com/sites/bernardmarr/2018/02/14/the-key-definitions-of-artificial-intelligence-ai-that-explain-its-importance/#3045e5854f5d> accessed 15 December 2019.
Ibid.
Tannya D Jajal, ‘Distinguishing between Narrow AI, General AI and Super AI’ Medium (21 May 2018) <https://medium.com/@tjajal/distinguishing-between-narrow-ai-general-ai-and-super-ai-a4bc44172e22> accessed 15 December 2019.
Ibid.
Ibid.
Compare below under the heading ‘Preventing AI from Taking over Comparative Law Work’.
Kamila Hankiewicz, ‘What Is the Real Difference between Automation and AI?’ Medium (10 August 2018) <https://becominghuman.ai/what-is-the-real-difference-between-automation-and-ai-366513e0c910> accessed 15 December 2019; Dave Evans and Access Planit, ‘The Fundamental Differences between Automation and AI’ VentureBeat (4 October 2017) <https://venturebeat.com/2017/10/04/the-fundamental-differences-between-automation-and-ai/> accessed 15 December 2019.
Mikell P Groover, ‘Automation’ Encyclopedia Britannica (22 March 2019) <https://www.britannica.com/technology/automation> accessed 15 December 2019.
Evans and Planit (n 18).
Shivam Singh Sengar, ‘What Is the Difference between Automation and Artificial Intelligence?’ Quora (4 May 2016) <https://www.quora.com/What-is-the-difference-between-Automation-and-Artificial-Intelligence>.
Christine N (n 2) 4.
Palace (n 2) 218–19: ‘This is the story of AlphaGo Zero, the first artificial intelligence to learn tabula rasa—meaning from a “clean slate” without any human input. AlphaGo Zero started out it with no prior knowledge of the game Go except for its rules. Nonetheless, after forty days of playing against itself, it outperformed the algorithm that defeated the world’s best human Go player’; cf Denicola (n 2) 253.
Tromans (n 1) 3.
Lee Kai-fu (n 2); compare, however, that which is critical in relation to AI’s ability to conduct complex legal analysis. Paliwala (n 4) 108.
Tromans (n 1) 3; for (older) sceptical statements regarding the use of natural language processing (NLP), cf Tamsin Maxwell and Burkhard Schafer, ‘Natural Language Processing and Query Expansion in Legal Information Retrieval: Challenges and a Response’ (2010) 24(1) Intl Rev L, Computers & Technology 63, 65; Ben Goertzel and Cassio Pennachin, ‘The Novamente Artificial Intelligence Engine’ in Ben Goertzel and Cassio Pennachin (eds), Artificial General Intelligence (Springer 2007) 63, 122: ‘There is a host of natural language processing (NLP) technology out there, but the plan fact is that none of it works very well. Modern NLP technology works passably well if one of two criteria is met; the sentences involved are simple, or the sentences involved all pertain to a single, very narrow domain’; Celeste Tito, ‘Artificial Intelligence: Can Computers Understand Why Two Legal Cases are Similar?’ (1987) 7 Computer/Law J 409, 426.
Mauro Bussani and Ugo Mattei, ‘Diapositives versus Movies: The Inner Dynamics of the Law and Its Comparative Account’ in Mauro Bussani and Ugo Mattei (eds), The Cambridge Companion to Comparative Law (CUP 2012) 3.
Zweigert and Kötz (n 7) 2.
Lutz-Christian Wolff, ‘Comparing Chinese Law…But with Which Legal Systems’ (2018) 6(2) CJCL 151, 153; cf WJ Kamba, ‘Comparative Law: A Theoretical Framework’ (1974) 23 ICLQ 485, 505–10; Esin Örücü, ‘Developing Comparative Law’ in Esin Örücü and David Nelken (eds), Comparative Law: A Handbook (Hart Publishing 2007) 43, 56–62.
Cf Wolff (n 29).
This section is based on Wolff (n 29). Zweigert and Kötz (n 7) 33 Wolff (n 29) 153; Mark Van Hoecke, ‘Methodology of Comparative Legal Research?’ (2015) 12 Law and Method 1, para 1 <https://www.bjutijdschriften.nl/tijdschrift/lawandmethod/2015/12/RENM-D-14-00001> accessed 15 December 2019: ‘Researchers get easily lost when embarking on comparative law research. The main reason being that there is no agreement on the kind of methodology to be followed, nor even on the methodologies that could be followed.’
Wolff (n 29) 154: ‘(L)egal research methodology aims to prescribe the most effective way forward and, thus, the way forward that is most promising in terms of avoiding mistakes and generating useful outputs. The identification of goal(s) of legal research is the first step in this regard. When compared with goalless and, thus, random legal work, goal-oriented and, thus, targeted research has consequently a much higher chance of success’; Jaakko Husa, ‘Methodology of Comparative Law Today: From Paradoxes to Flexibility?’ (2006) 8(4) Revue Internationale de Droit Comparé 1095, 1096: ‘Method is here understood to be an orderly and systematic manner in which research is done and, in accord, methodology is the field that deals with questions concerning methods, in this case especially methods of comparative study of law’; cf Kamba (n 29) 489.
Lomio, Spang-Hanssen and Wilson (n 11) 65; Jaakko Husa, ‘A New Introduction to Comparative Law’ (Hart Publishing 2015) 98; see also John C Reitz, ‘How to Do Comparative Law’ (1998) 46 Am J Comp L 617, 624; Mathias Siems, Comparative Law (2nd edn, CUP 2018) 16.
Wolff (n 29) 154.
Cf Lomio, Spang-Hanssen and Wilson (n 11) 59–60; Marieke Oderkerk, ‘The Importance of Context: Selecting Legal Systems in Comparative Legal Research’ (2001) XLVIII Netherlands Intl L Rev 293, 312: ‘almost infinite’; Kamba (n 29) 490, 508; Reitz (n 33) 624; Mary Ann Glendon, Paolo G Carozza and Colin B Picker, ‘Comparative Legal Traditions: Text, Materials and Cases on Western Law’ (4th edn, West Publishing 2014) 15: ‘Discussions of the goals of comparative law often draw on overly sharp distinction between its practical and its scientific aims.’; Örücü (n 29) 4, 44: ‘Objectives are varied…even working towards furthering of world peace and tolerance have been attributed to comparative law.’; Van Hoecke (n 31) under ‘1. Why compare?’; Gerhard Dannemann, ‘Comparative Law: Study of Similarities or Differences?’ in Reimann and Zimmermann (n 7) 383, 401.
Kamba (n 29) 495–9; Lomio, Spang-Hanssen and Wilson (n 11) 64; cf Oderkerk (n 35) 312; Van Hoecke (n 31) under ‘1. Why compare?’; Reitz (n 33) 624.
Lomio, Spang-Hanssen and Wilson (n 11) 64 (‘understanding’); Oderkerk (n 35) 312 (‘reflection’); Kamba (n 29) 494 (‘academic understanding’); Hessel E Yntema, ‘Comparative Legal Research: Some Remarks on ‘Looking out of the Cave’ (1956) 54 Michigan L Rev 899, 901: (‘constant refinement and extension of our knowledge of law’); Reitz (n 33) 624; Zweigert and Kötz (n 7) 15: ‘The primary aim of comparative law, as of all sciences, is knowledge’; Rodolfo Sacco, ‘Legal Formants: A Dynamic Approach to Comparative Law’ (1991) 39 Am J Comp L 1, 4 (‘acquisition of knowledge’).
Jan M Smits, ‘Comparative Law and Its Influence on National Legal Systems’ in Reimann and Zimmermann (n 7) 519: ‘When the influence of comparative law on national courts is discussed, it is often the voluntary use of foreign law in purely domestic disputes that forms the centre of attention. There are now examples of such influence in almost every legal system, even though important differences between various countries are to be appreciated.’
Cf Kamba (n 29) 499; Glendon, Carozza and Picker (n 35) 15; Zweigert and Kötz (n 7) 16: ‘Legislators all over the world have found that on many matters of good laws cannot be produced without the assistance of comparative law, whether in the form of general studies or reports prepared on the topic in question’; Smits (n 38) 513.
Kamba (n 29) 490–4, lists the ‘teaching and study of law’ as the first goal; Örücü (n 29) 44.
Cf Örücü (n 7) 447 (‘macro-level, meso-level and micro-level comparison’).
Cf Kamba (n 29) 505–10; Örücü (n 29) 56–62; Van Hoecke (n 31) under ‘3. What has to be compared?’.
Van Hoecke (n 31) under ‘4.1. The functional method’.
See text accompanying note 36.
Cf Wolff (n 29) 153–73.
Cf Aleksandar Momirov and Andria Naudé Fourie, ‘Vertical Comparative Law Methods: Tools for Conceptualising the International Rule of Law’ (2009) 2 Erasmus L Rev 291.
Wolff (n 29) 152.
Ibid. Comparative legal history work related to the same jurisdiction is normally not regarded as falling within the scope of comparative law. Zweigert and Kötz (n 7) 2; James Gordley, ‘Comparative Law and Legal History’ in Reimann and Zimmermann (n 7) 744, 754; Heikki Pihlajamäki, ‘Merging Comparative Law and Legal History: Towards and Integrated Discipline’ (2018) 66 (4) Am J Comp L 733; Eltjo Schrage and Viola Heutger, ‘Legal History and Comparative Law’ in Smits (n 7) 393, 405: ‘It is not always easy to draw the borderline between the two disciplines.’
Where the object of a (macro-)comparative study are legal systems as such, the choice of the jurisdictions to be compared corresponds with the selection of the object of the comparative study.
Cf Oderkerk (n 35) 307; Kamba (n 29) 510.
See text accompanying note 36.
Cf Zweigert and Kötz (n 7) 41; Kamba (n 29) 489; Van Hoecke (n 31) under ‘2. Choice of legal systems to be compared’; Carlo Garbarino, ‘An Evolutionary Approach to Comparative Taxation: Methods and Agenda for Research’ (2009) Am J Comp L 677, 690.
Cf Oderkerk (n 35) 317.
Wolff (n 29) 152.
For a more comprehensive discussion see ibid 158–64.
Oderkerk (n 35) 305; Kamba (n 29) 508, comments on the required qualification of comparatists: ‘It seems obvious that the making of an intelligent selection of legal systems demands on the part of the comparatist a good conspectus or panoramic view of the legal systems of the world. He must be grounded in the bases or criteria for their classification and must have a good knowledge of the characteristics of the major groups or families of systems.’
Cf Van Hoecke (n 31) under ‘2. Choice of legal systems to be compared’; Gerber (n 7) 722. It is of course possible that research goals are identified on the basis of personal features. However, in cases of this kind the scope of respective research may be limited, which has to be acknowledged and justified. Cf Wolff (n 29) 159.
Wolff (n 29).
Cf Glendon, Carozza and Picker (n 35) 33–4; Zweigert and Kötz (n 7) 63–73; Mariana Pargendler, ‘The Rise and Decline of Legal Families’ (2012) 60 Am J Comp L 1043–74; H Patrick Glenn, ‘Comparative Legal Families and Comparative Legal Traditions’ in Reimann and Zimmermann (n 7) 421.
For the fact that normally the so-called parent system as opposed to affiliated legal systems should serve as representative system cf Zweigert and Kötz (n 7) 41; Oderkerk (n 35) 300; for the argument that not only parent legal systems produce original legal solutions while this could as a matter of principle also be achieved at the level of affiliated legal systems, see Zweigert and Kötz (n 7) 42; Oderkerk (n 35) 302, both with reference to Ulrich Drobnig, ‘Methodenfragen der Rechtsvergleichung im Lichte der International Encyclopedia of Comparative Law’ in Ernst von Caemmerer and Soia Mentschikoff and Konrad Zweigert (eds), Ius privatum gentium: Festschrift für Max Rheinstein zum 70. Geburtstag am 5. Juli 1969, vol 1 (Ius privatum gentium: Festschrift for Max Rheinstein for his 70th Birthday on 5 July 1969) (JCB Mohr 1969) 221. For Drobnig’s suggestion that there is only a limited number of solution types for any legal problem (so-called Lösungstypen) and that legal systems that represent these solution types should consequently be compared, see Oderkerk (n 35) 303, 317; also cf Örücü (n 7) 442, 444: ‘This approach can also be called the universalist approach to human needs which from the belief that social problems are universal, the laws respond to these needs in various ways, but that the end results are comparable.’
Cf Zweigert and Kötz (n 7) 63, 67; Siems (n 33) 94.
Horatia Muir Watt, ‘Globalization and Comparative Law’ in Reimann and Zimmermann (n 7) 579, 579; William Twining, ‘Globalization and Comparative Law’ in Örücü and Nelken (n 29) 69, 69–75.
Wolff (n 29) 159–60.
Cf Siems (n 33) 16.
Cf Kamba (n 29) 508; Jaques du Plessis, ‘Comparative Law and the Study of Mixed Legal Systems’ in Reimann and Zimmermann (n 7) 477, 478; Örücü (n 7) 442: ‘(T)here is nothing in the logic of comparative inquiry dictating that comparison be limited to any specific level or unit.’
Wolff (n 29) 161.
Cf Kamba (n 29) 501; John Bell, ‘Legal Research and the Distinctiveness of Comparative Law’ in Mark Van Hoecke (ed), Methodologies of Legal Research: Which Kind of Method for What Kind of Discipline? (Hart Publishing 2011) 155, 157: ‘The objective of many comparative lawyers has been to achieve harmonization if not unification’; Van Hoecke (n 31) under ‘1. Why compare?’.
Wolff (n 29) 161–2.
Örücü (n 7) 442; cf Van Hoecke (n 31) under ‘6. Tertium comparationis’.
Ibid.
Örücü (n 7) 443
Siems (n 33) 14, 25–8; Örücü (n 7); Zweigert and Kötz (n 7) 34; critical of the notion of a functional method, see Ralf Michaels, ‘The Functional Method of Comparative Law’ in Reimann and Zimmermann (n 7) 339–82; Jaakko Husa, ‘Functional Method in Comparative Law: Much Ado about Nothing?’ (2013) 2(1) Eur Property LJ 4.
Van Hoecke (n 31) under ‘4. How does one compare?’, who distinguishes between the functional, structural, analytical, the law-in context, historical, and the common core methods while emphasizing that these methods ‘are not mutually exclusive. It is even possible to combine all of them in and the same research. The name of the method points to the specific nature of that approach, without excluding its combination with another method.
Lin Siyi, ‘The Goals of the Law of Unjust Enrichment’ PhD dissertation, Chinese University of Hong Kong (2019) under ‘Chapter 1 Introduction’ ‘5.2 Doctrinal Research’; Terry Hutchinson and Nigel Duncan, ‘Defining and Describing What We Do: Doctrinal Legal Research’ (2012) 17 Deakin L Rev 83, 101.
Cf Bell (n 67) 170: ‘In short, comparative lawyers cannot be content to present rules without some reference to the organizational setting, the procedural context and the conceptual structure within which legal problems emerge and the rules are operated. It is in this setting that it is possible to work out the extent of differences or similarities between the systems.’; also see below at ‘Deducing Value Judgements and Considering Path Dependence. F.’
Bell (n 67).
University of Leicester, ‘8.3.1 Empirical research’ <https://www.le.ac.uk/oerresources/criminology/msc/unit8/page_04.htm> accessed 15 December 2019; also compare for ‘numerical comparative law’, where ‘not only empirical data (such as the number of courts and their ability to enforce the law) but also the ‘law as such’ (ie statutes and case law) are translated into numbers’. Siems (n 33) 146–87; Mathias M Siems, ‘Numerical Comparative Law: Do We Need Statistical Evidence in Law in Order to Reduce Complexity?’ (2005) 13 CJICL 521.
Van Hoecke (n 31) under ‘4.5 The historical method’ (‘just one part of the ‘law-in-context method’).
Compare the discussion above under the heading ‘Artificial Intelligence’.
Christine N (n 2) 4.
Stephen Hawking, quoted by Arjun Kharpal, ‘Stephen Hawking Says A.I. Could be “Worst Event in the History of Our Civilization”’ CNBC (6 November 2017) <https://www.cnbc.com/2017/11/06/stephen-hawking-ai-could-be-worst-event-in-civilization.html> accessed 15 December 2019.
See section below under the heading ‘Preventing AI from Taking Over Comparative Law Work’.
See discussion above.
Ibid.
Ibid.
Ibid; see, however, Jansen (n 7) 314.
Cf Jansen (n 7) 306; Reitz (n 33) 624, 628–31; Zweigert and Kötz (n 7) 40: ‘The question how the comparatist should set out and how far he should go in his search for material is intimately related to the meaning and purpose of comparative law, to its very methods and thoughts’; Örücü (n 7) 446–8.
Zweigert and Kötz (n 7) 37: ‘The comparatist who wants to find in a foreign system the rules which are functionally equivalent to those which interest him in his native law requires both imagination and discipline.’
Compare the remarks made 10 years ago by Maxwell and Schafer (n 26) 65: ‘Lack of available context in queries and documents can limit NLP-related gains because there is minimal linguistic information to extract. …This gets to the heart of what is most difficult about language processing. Language is complex, unpredictable and productive, and accounting for phenomena such as ungrammatical sentences, disfluencies, negation, multiword terms and polysemy can be difficult’; see also Richard E Susskind, ‘Artificial Intelligence, Expert Systems and Law’ (1990) 5 Denning LJ 105, 107.
Tromans (n 1) 4.
Ibid.
See section above under the heading ‘Artificial Intelligence’.
Tromans (n 1) 4.
Cf Reitz (n 33) 618: ‘I wish to insist that the comparative method involves explicit comparison of aspects of two or more legal systems. Some may object that any description of foreign law is implicitly comparative because all descriptions of foreign law are at a minimum trying to make the law of one comprehensible for those trained in a different system. But I reject that argument on the grounds that the step of actually drawing the comparison is crucial to realizing the intellectual benefits of comparison.’
Cambridge Dictionary (CUP 2019): ‘comparing’ <https://dictionary.cambridge.org/dictionary/english/comparing> accessed 15 December 2019.
Cf Jansen (n 7) 306 (‘similarities and dissimilarities’); Reitz (n 33) 620; Dannemann (n 35) 418; Örücü (n 7) 448 (‘identification or discernment of differences and similarities between the phenomena under comparison’).
Tromans (n 1) 4.
Zweigert and Kötz (n 7) 46 (‘the comparatist must proceed to a critical evaluation’); Örücü (n 7) 449: ‘[C]omparative inquiry should not end at description, but move on into explanation where the real comparison starts, and then, on into confirmation of findings.’
Reitz (n 33) 625: ‘The real power of comparative analysis arises precisely from the fact that the process of comparing ‘apples’ and ‘oranges’ forces the comparatist to develop constructs like “fruit.”’
Jansen (n 7) 312–14.
Franklin (n 4) 245.
Compare discussion above under the heading ‘Introduction’.
Jansen (n 7) 314: ‘(C)omparisons are always related to a tertium comparationis that is often implicit in the conceptual structure chosen. …Such tertia comparationis are not objectively ‘in the air’; rather they result from a choice about ‘what matters’, that is, which aspects of the law are relevant for the comparative lawyer, and which aspects of the law might benefit from the additional knowledge which comparison provides. …[A]lthough comparisons can be described as normatively neutral, it is misleading to suggest that evaluations come into play only after the process of comparison is finished.’
Cf Jansen (n 7) 307–8, 314; for unification with prior harmonization, see Örücü (n 29) 450 (‘political choices’); Van Hoecke (n 31) under ‘5.5.2 With varying application according to local culture’.
From the methodological point of view, cf Örücü (n 7) 446.
Jansen (n 7) 306: ‘[L]egal rules and legal texts are typically deeply rooted within a specific, economic, political, moral, and cultural background, which can often only be explained from a historical perspective’; Reitz (n 33) 625: ‘The argument for domestic law reform has to be made in terms of normative claims acceptable within the domestic legal system’; cf Örücü (n 7) 445, 449; Van Hoecke (n 31) under ‘4.4 The law-in-context method’.
See section under the heading ‘Setting Methodological Parameters’; Wolff (n 29) 164.
See section above under the heading ‘Collecting Data and Other Information’.
Cf Timo Rademacher, ‘Wenn neue Technologien altes Recht durchsetzen: Dürfen wir es unmöglich machen, rechtswidrig zu handeln?’ (When new technologies enforce old law: are we free to make illegal actions impossible?) (2019) 74 Juristische Schulung 702, 707.
Compare section above under the heading ‘Artificial Intelligence’.
Paliwala (n 4) 108: ‘However, when information technology has been applied to deeper legal processes, which involve the very nature of law, the result has not been very successful. This is especially so in relation to the application of AI systems to law.’ Paliwala (n 4) suggests that the problems may have little to do with the limits of AI technology, but with the fact that ‘[m]any involved with AI and law still refuse to acknowledge that there are underlying problems with the way they conceptualize the nature of legal reasoning.’
Storrs Hall (n 6) 250; cf Denicola (n 2) 256: ‘Combined results from surveys of artificial intelligence experts estimate a 50% chance of human-level machine intelligence by 2040 and a 90% probability by 2075’; Paliwala (n 4) 112 (‘quite possible’); Jajal (n 14), at ‘A Melding of Humans and Machines’.
See section above under the heading ‘Deducing Value Judgements and Considering Path Dependence’.
Örücü (n 7) 443.
Without AI reference from the functional point of view, cf ibid.
For so-called ‘impossibility structures’ that, through technological means, prevent the violation of legal rules, cf Michael L Rich, ‘Should We Make Crime Impossible?’ (2013) 36 Harv JL & Public Pol’y 795.
Compare section above under the heading ‘Setting Methodological Parameters’.
Ibid.
Ibid.
See section above under the heading ‘Artificial Intelligence’.
Hawking (n 81); also cf Elon Musk, quoted by Luciano Floridi, ‘Should We Be Afraid of AI?’ AEON (9 May 2016) <https://aeon.co/essays/true-ai-is-both-logically-possible-and-utterly-implausible> accessed 15 December 2019: ‘We should be careful with artificial intelligence’; Palace (n 2) 241.
Cf ‘How to Do Some Restrictions on Artificial Intelligence in the Future?’ Ready for AI (28 January 2019 ) <https://readyforai.com/article/how-to-do-some-restrictions-on-artificial-intelligence-in-the-future/> accessed 15 December 2019, under ‘beyond cognition’.
Compare the example quoted by Floridi (n 121): ‘Microsoft introduced Tay—an AI-based chat robot—to Twitter. They had to remove it only 16 hours later. It was supposed to become increasingly smarter as it interacted with humans. Instead, it quickly became an evil Hitler-loving, Holocaust-denying, incestual-sex-promoting, “Bush did 9/11”-proclaiming chatterbox. Why? Because it worked no better than kitchen paper, absorbing and being shaped by the nasty messages sent to it. Microsoft apologised’; David Lee, ‘Tay: Microsoft Issues Apology over Racist Chatbot Fiasco, BBC News (25 March 2016) <https://www.bbc.com/news/technology-35902104> accessed 15 December 2019.
Floridi (n 121).
Christine N (n 2) 04.
Julia Bossmann, ‘Top 9 Ethical Issues in Artificial Intelligence’ World Economic Forum (21 October 2016) <https://www.weforum.org/agenda/2016/10/top-10-ethical-issues-in-artificial-intelligence/> accessed 15 December 2019.
Storrs Hall (n 6) 250; cf Denicola (n 2) 256; Jajal (n 14), at ‘A Melding of Humans and Machines’.
