The Impacts of Artificial Intelligence on Research in the Legal Profession

Legal research is an indispensable skill for lawyers. Therefore, it is always necessary for lawyers to engage in legal research in due course of trying to alleviate various legal problems. Although the purpose and methodology of the research may vary from lawyer to lawyer, doing research is a common activity. As a result, the quest to assess the impacts of artificial intelligence (hereinafter ‘AI’) on legal research allows one to measure the influence of AI on the legal profession in general. Moreover, with the advent of Legal AI, it is now evident that the legal profession is not immune from disruption. According to the above, this article discusses the impacts of AI on research in the legal profession in general in accomplishing various lawyerly tasks by different legal professionals.


The Historical Background of Artificial Intelligence
In 1950, British mathematician Alan Turing published a paper on computing machinery and intelligence posing the question of whether machines can think. 13  they should accomplish the specific task. These programs use 'iteration', a process of repetitively feeding data into an algorithm to improve their outputs. Over time, these programs can make their judgments based on previous data from similar but not identical tasks. 39 On the other hand, 'deep learning' is a technique within machine learning tools that aims to enable example-based learning of machines and autonomous systems. 40 Instead of instructing the system with a set of pre-determined instructions, deep learning provides a model for the machine to evaluate examples and infer patterns for the solving of future problems. It is from the harmonious application of the stated components that AI will be capable of processing a given instruction and provide a required outcome. 41 Accordingly, AI is viewed as an artificial system that performs tasks under varying, but predictable circumstances and without significant human oversight. Such systems could also learn from their experiences while improving their performances (for the future) and might even solve tasks requiring human-like perception, cognition, planning, learning, communication, or physical actions. 42 On the other hand, given its elusive nature, different professionals have defined AI differently from different perspectives. This made it difficult to coin a universally agreeable definition of AI. For example, from the perspective of what it is made up of, AI can be portrayed as a system that includes both hardware and software components, which may refer to a robot, a program running on a single computer, a program run on networked computers, or any other set of components that hosts an AI. 43 From the viewpoint of legal recognition, AI is most often considered as work resulting from intellectual activity that can be protected by intellectual property law as software 44 45 Under certain conditions, AI can also be protected by a software patent. 46 The law also accords protection to AI systems that are inseparably incorporated into physical devices such as robots by considering them as products. 47 From the perspective of its end purpose, AI can be defined as the process of simulating human intelligence through machine processes. 48 In this regard, the end goal is to create artificially intelligent machines, often in the form of robots that can perform traditionally human tasks better and more efficiently than humans ever could. 49 Accordingly, experts in the field of AI classify such artificially intelligent machines into two major types: the first is General AI that refers to an extremely complex machine (algorithm or set of algorithms) that think like people across multifaceted problem domains and have the ability to reason generally, which is the goal for the future (currently hypothetical). 50 The second is called Narrow (applied/ specialized) AI that refers to systems designed to execute specific tasks or a single function, and will never rival the cognitive depth of a human being. 51 Good examples are playing 'chess' or 'Go', or diagnosing an illness. Narrow AI is already functional in various aspects of human life often with greater accuracy and efficiency than human beings. 52 It should be noted that these two approaches to AI rely on machine learning, which is the process of teaching a program to learn from user-fed data to 45  respond to completely new data in the future, without the need to program a specific set of instructions for every possible data point. 53 From the perspective of its artificial nature, AI can be defined as a non-biological autonomous entity. However, the term autonomous in this definition should be taken as the ability of AI to process data by itself and by no means prohibits any situation in which human and AI experts are working alongside one another (also called co-robotics). 54 On the other hand, the notion AI is made up of two words "Artificial" which implies a good made by people, often as a copy of something natural, and "Intelligence" which may refer to: "the ability to learn and understand or to deal with new or trying situations", or "the skilled use of reason", or "the ability to apply knowledge to manipulate one's environment or to think abstractly as measured by objective criteria". 55 Accordingly, AI is an artificially developed intelligence, created as an alternative to humans, or a crafted machine with embedded learning and analysis capabilities, mastered to comply with real-life situations and to perform, as much as accurately possible, the tasks and works once done by men. 56 Simply put, AI is that activity devoted to making machines intelligent, and intelligence is that quality that enables an entity to function appropriately and with foresight in its environment.
As famously stated by, the British mathematician Alan Turing, one of the pioneers in the field of AI "a computer would deserve to be called intelligent if it could deceive a human into believing that it was human". 57 However, it should be noted that AI is not intelligent in the sense that it does not know what it is doing, or why it is doing it. An AI system is not really 'reasoning' or 'thinking' but is 53 Sean Semmler and Zeeve Rose, supra note 39 54  following a set of pre-programmed computational steps (expert systems) or mathematically analyzing a huge amount of data to infer a probability (machine learning). 58 As adequately emphasized by Steven Pinker, AI does not have intentionality or a real attitude, but only sets tasks and goals; unlike humans, it does not make real judgments based on principles, rules, priorities, or values. 59 In 1956, Professor John McCarthy, at a conference held in Dartmouth College, New Hampshire, USA was credited with introducing the term 'AI' as: 'the science and engineering of making intelligent machines, especially intelligent computer programs. 60 Since then AI is portrayed as a machine that behaves in ways that would be called intelligent if a human were so behaving. 61 Similarly, in 1968 Marvin Minsky, one of the founders of AI, described AI as the science of making machines do things that would require intelligence if done by man. In this sense, AI is pursued at least for two reasons: to understand the workings of human intelligence and to create useful computer programs and computers that can perform intelligently. 62 Therefore, AI can be broadly characterized as intelligence by machines and software. 63 Similarly, AI can be practically defined as the theory and development of computer systems that can perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages. 64 Thus, it is quite evident from the above definitions that human intelligence is taken as a benchmark to measure AI.
Similarly, Richard Susskind, a prominent expert in AI, defined AI as a field of study concerned with the design, development, and implementation of computer systems that can perform tasks and solve problems of a sort for which human intelligence is normally thought to be required. 65 So AI as a field of science and technology is employed where processes are used to carry out tasks, i.e. the processes that mimic, imitate, or simulate intelligence. 66 AI may also be defined by reference to the tasks it performs (such as visual perception, speech recognition, decision-making, and translation between languages) and the processes used to perform tasks: expert systems, machine learning (supervised, unsupervised, neural networks) and so on. 67 In general, AI is an umbrella term that refers to teaching a machine how to do a task that was thought to be human. 68 That is why, in its January 2018 book, 'The Future Computed' Microsoft defined AI as "a set of technologies that enable computers to perceive, learn, reason and assist in decision-making to solve problems in ways that are similar to what people do." 69 When it comes to AI in the legal profession, it can be conventionally defined as programing computer technologies (such as machine learning, natural language processing, speech recognition, legal robotics, planning, natural image understanding, rule-based expert system, neural networks, logic programming, artificial vision, machine learning, and neural networks) to process, analyze and finalize various legal tasks historically performed by lawyers. 70 In summary, AI collectively covers a range of technologies from simple software to sentient robots, and everything in between and unavoidably includes both algorithms and data. 71

Definition of Legal Research
Legal research can be defined differently from different perspectives. The term is composed of two words, 'Legal' and 'Research'. Accordingly, the term 'research' refers to any gathering of data, information, and facts for the advancement of knowledge. 72 Similarly, research can be defined as a studious inquiry or examination; especially investigation or experimentation aimed at the discovery and interpretation of facts, revision of accepted theories or laws in the light of new facts, or practical application of such new or revised theories or laws. 73 Therefore, research is the act of searching into a matter closely and carefully, inquiry directed to the discovery of truth and in particular, the trained scientific investigation of the principles and facts of any subject, based on an original and first-hand study of authorities or experiment. 74 On the other hand, research can also be portrayed as a process of steps used to collect and analyze information to increase an understanding of a topic or an issue. 75 Research refers to the process of identification of a problem, the ascertainment of the relevant facts, their logical ordering and classification, the use of logic (science) to interpret the collected and classified facts, and the assertion of conclusions premised on and supported by the collected information. 76 In this sense, research is a creative and systematic work undertaken to increase the stock of knowledge and to devise new applications of available (existing) knowledge. 77 It is logical to conclude from the above definitions that 'research' is the careful, diligent, exhaustive, and systematic (scientific) investigation (search/pursuit) of a specific subject matter (knowledge) to know (discovering) the truth and making an original contribution in the existing stock (body) of knowledge. On the other hand, legal research can be defined as a systematic finding or ascertaining of the law on the identified problem or in the given area as well as an inquiry into law to solve a particular problem or making advancement in the science of law. 78 Accordingly, in a rough sense, legal research can be described as the process by which lawyers identify (find), read (retrieve), interpret (analyze and synthesize) the content of the law, and explain the law to their clients or to judges or to support legal decision making. 79 Therefore, legal research is an important part of being a lawyer, and that is why it is believed that the clients are indeed paying a price for the quality of legal research undertaken by lawyers on a particular problem. 80 However, it should be emphasized that as far as they can access the law and have the required knowledge, non-lawyers can also do legal research, whether to resolve a private dispute, understand and work with a legal professional, or assist in their own academic or professional development and so on. 81 Legal research also includes the process of identifying the pertinent facts and legal issues related to a particular problem, finding and using relevant secondary sources, finding and using governing (appropriate and up to date) primary sources (law) and case law, analyzing the law as it relates to the legal issues and the facts of the case (application) and communicating the findings of the inquiry and analysis. 82 On the other hand, the fact that the scope of the law is vast and its nature is ever-changing, and it is different from jurisdiction to jurisdiction, has made it impossible for anyone to know all the law about every topic and from every jurisdiction. 83 In turn, these facts made legal research indispensable for lawyers to be able to stay abreast of the continual changes in the law while making proper (relevant) representations of the law to courts as it relates to their clients' cases. Accordingly, legal research is the process by which lawyers find the law, cut 78  how they use a 'rule' and arrive at a particular decision. It is via the underway of pertinent research, logical deduction, and legal reasoning that a Judge injects 'life' into 'law'. 88 Practicing lawyers are the other groups that undertake legal research daily and as an exercise of their occupation. Research is an intuitive aspect of legal work. 89 Legal research skills have been identified as a core skill for lawyers. Good legal research skills are a necessary step in attaining the ability to think like a lawyer and achieving valid legal reasoning outcomes. 90 Practicing lawyers, as a professional, have to advise their clients and plead cases on their behalf in the court of law. They are also required to give legal opinions and advice on issues referred to them by clients. Therefore, it is as part of the requirement of the office and profession that lawyers need to undertake systematic research for 'finding' the law and thereby provide a solution to legal problems. 91 By far, legal academicians are the ones that have a predominant association with legal research. For example, law professors are required by their universities to undertake legal research as a part of their professional commitment. 92 In addition, law students must undertake various types of researches in fulfillment of their law degree. Therefore, strong legal research and writing skills are fundamental tools of legal scholarship. 93 Doing legal research is an integral part of everyday teaching in academic institutions. Professors need to study the law and update themselves with relevant knowledge daily to serve as reliable sources of knowledge to their students. In addition, various intellectual challenges, academic reputation, and requirements of academic ranks and degrees are various reinforcements for academia to engage in legal research. 94

The Positive Impacts of AI on Research in the Legal Profession
Legal research in particular and the legal profession in general, are not immune from disruption by AI. Susskind has successfully predicted in 2013 that AI technologies will bring radical change in the legal profession in the next ten years. 95 Susskind further argued that it is simply inconceivable that information technology will radically alter all corners of the economy and society and yet somehow legal work will be exempt from any change. 96 Previously, because lawyers are highly trained and skilled professionals who identify the legal issues, gather the relevant facts and determine the likely outcome of a court decision to adjudicate a dispute by exercising judgment, using their experience and intuition, to assess the merits of a case to determine the best way to proceed, it was generally believed that lawyerly tasks could only be performed by highly-skilled professionals. 97 However, recent developments in AI have challenged the traditional conceptions and proved that legal practice is not immune from AI. As a result, AI has and will cause a great disruption in legal research in particular and legal practice in general. 98 Accordingly, recent developments in AI, such as natural-language processing and machine learning have challenged the traditional conceptions of human lawyer expertise. Various complex tasks that used to require human effort have been automated in ways that reduce cost and offer greater accuracy and precision, which is a good indicator that legal practice is not immune from these technological advances. That is why machine intelligence will cause a great disruption in the market for legal services regarding discovery, legal search, document generation, brief generation, and prediction of case outcomes. 99 In this regard, AI provides celerity, simplicity, and effectiveness in solving a multitude of legal problems by researchers. More automation reduces transaction costs dramatically, 95  AI can also perform automated tasks and adopt mass decisions efficiently. The use of AI is critical in legal research in terms of efficiency in searching, classifying, filtering, rating, and ranking issues, facts, ideas, laws, and so on. 101 On the other hand, AI combined with computer systems is also capable of many other impressive feats that make the undertaking of legal research very easy. Such as recognizing and pointing out spelling errors and finding bad writing, and suggesting the rewriting of bad sentences. 102 Moreover, the weak version of AI is already serving as a large improvement on existing legal research tools such as Lexis and Westlaw, to assemble an array of relevant cases, suggest similarities and differences, and sketch arguments and counterarguments. On the strong version, however, in the future, AI will help lawyers or even judges, to engage in legal reasoning in researching the most relevant cases or laws to solve particular legal problems. 103 AI is also a very useful tool for law and legal science in general. By applying knowledge to find a solution to legal problems, AI applications are assisting in legal reasoning. AI provides tools and techniques developed to solve specific problems in law in general. Legal science recognizes the usefulness of AI for legal reasoning and research. Legal reasoning is a general concept that refers to a process of forming and providing a justifiable answer to a particular legal question. 104 For example, by searching databases of legal texts and identifying which cases are relevant to the respective ongoing judicial proceedings. 105  Besides, not a few numbers of technologies assist lawyers in the due course of legal research, such as in identifying problematic clauses in contracts or planning a winning strategy in intellectual property lawsuits. For instance, another area of application for AI in the legal field is online dispute resolution, which is destined to solve disagreements between parties that entered into a contract via an electronic platform. 108 The ability of AI to analyze vast amounts of data is also used, for example, in digital forensics. AI is also used for predictions, such as for determining which crime scenes will offer the best opportunity of recovering a forensic sample. 109 Likewise, in human rights law practice and research, AI has offered an improved ability to monitor and document war crimes and human rights abuses. AI in the 21 st century has ushered in the golden age of surveillance by states, corporations, and non-state actors. Human putting a significant strain on how the judiciary goes about doing things. 113 Moreover, the Chief Justice when asked whether he could foresee a day when AI would assist with courtroom fact-finding or judicial decision-making affirmed that 'It's a day that's here,' and AI is putting a significant strain on how the judiciary goes about doing things.' 114 Courts are also utilizing AI in making judicial decisions. Courts in the USA utilize advanced algorithms to assist in pretrial detainment of the accused. For example, the 'Public Safety Assessment tool is utilized in 29 American jurisdictions to determine the risk associated with defendants. 115 On the other hand, acquiring legal representation to take and defend a case in court differs in various countries and can be a tedious, lengthy and costly process. Robots are also providing the possibility to have a positive impact in several aspects of the processes of the judicial system, as automation outperforms humans and increases productivity. 116 Therefore, AI will have a positive impact in shortening the judicial process via automation and increased productivity. 117 Studies show that the utilization of AI in courts could result in up to a 13% decline in lawyers' hours that would enable more rapid processing of cases in courts. 118 Moreover, Legal AI is helping attorneys to become more efficient in research and serve a wider range of clients on a broader range of issues. If anything, legal AI is allowing lawyers to perform more work, with less effort, and more money. Accordingly, the only lawyers with anything to fear are those who refuse to embrace change for AI has the potential to break into almost every aspect of legal practice. 119 Moreover, AI is influencing legal research and practice by making lawyers more efficient in their job, automating legal services, and updating the law itself. AI is challenging traditional legal concepts by forcing the law to adapt to new developments in technology. Concurrently, the law will be shaping developments in AI by imposing new standards, guidelines, as well as 113  Currently, most of the work of associates is geared towards legal research and due diligence that is highly amenable to be assisted by AI tools, which made modern lawyers more effective than the traditional approach. 122 That is why emerging legal tech companies allow these associates to use AI capabilities to identify legal authorities relevant to particular questions, which made them more effective than the traditional, labor-intensive approach utilized by most big firms today. Accordingly, researches show that law firms that use AI tools: (1) have better information retrieval quality, (2) are intuitive to use requiring little training, and (3) will drastically cut working hours. This enabled firms to abstain from hiring many associates and spend less time on research, which freed associates for other substantive activities by automating legal grunt work. 123 Currently, due to the advent of AI support, firms will no longer need to hire many associates to sift through contracts and conduct legal research. The use of AI tools is helping to maximize the efficiency of each research project, forcing firms to either cut down on hiring or put their associates to better use. Furthermore, with legal grunt work becoming automated, associates will be free to engage in more substantive work at earlier stages in their careers. 124 Similarly, it is evident that the Big Law Firm model will disappear soon due to developments in legal AI. This is because AI will create universal access to services that previously could only be accomplished by teams of highly educated attorneys. 125  than smaller firms do, which enables them to attract new clients while retaining their old clientele, which will be discouraging for smaller firms to join the market. 127 Hence, the use of AI in legal research and practice is inevitable due to its competitive, comparative, and differential advantages. AI tools allow law firms to reduce the labor hours required for research and spend more time on high-value legal matters, which enables the firm to produce cheaper services while attracting more customers thereby creating a competitive advantage. The use of AI tools is also enabling firms to attract both curious On the other hand, even on an individual level, AI is enabling lawyers to do more work at a given time, which has increased their efficiency. With the ability to work efficiently, lawyers are less tethered to work in large firms and still can perform capably. Efficiency will also empower lawyers to broaden their areas of specialization. At present, lawyers are using AI tools to maintain areas of expertise and develop new ones. 131 On the other hand, due to the fierce competition and cheap services brought by the advent of AI clients are becoming less willing to pay big prices for legal research. Clients are starting to demand fixed fees for work that was traditionally billed by the hour, which is forcing firms to lower their prices. 132 Additionally, with the increased availability of AI tools, client expectations could change in that they become less willing to pay six-figure bills for legal research by lawyers. Currently, Clients are starting to demand fixed fees for work that was traditionally billed by the hour and 127  it is also common for clients to demand that associate work not be included in their bill.
Clients are beginning to expect more value for their money. As a result, with these changing client expectations, firms must lower their prices and adapt, or otherwise, they will lose huge amounts of business. 133 On the other hand, AI also increased the availability of legal services to consumers without hiring an attorney. AI by taking over some lawyer activities has enhanced access to justice and enabling mass-scale representation. For example, in April 2016, the 'DoNotPay' robot had helped people overturn 160,000 of 250,000 parking tickets. Since its launch, this robot scored a success rate of 64 % appealing over $ 4 million. 134 Similarly, AI has also significantly transformed the ability to store and access legal information and it creates full access to legal information, which is a core skill of legal research by lawyers. It has extensively transformed how laws were maintained, learned, and   177 Researches show that TAR can yield more accurate results than an exhaustive manual review with much lower effort. 178 There are also adequate studies that show the benefits of e-discovery, which can amount to saving 70% or more time. 179

The Negative Impacts of AI on Research in the Legal Profession
Accordingly, studies show that AI cannot read legal texts like lawyers can; applications can only extract some meaning from legal texts; machine language yields answers but not explanations; AI cannot usually explain its answers to legal questions; Question and Answer systems do not understand legal reasoning; 186 a tool cannot also reason about how different circumstances would affect its answers, and the majority of the AI tools also cannot work with total independence from human support. 187 On the other hand, according to Susskind, several problems and obstacles have been faced and will continue to confront the development of legal AI such as the lack of knowledge engineers, the lack of domain experts, the lack of existing methodology to be used by designers for the development of expert systems in the legal area, the lack of adequate AI tools, difficulties in quality control of legal AI systems, and the presence of huge concerns on the legal implications of AI tools. 188 Moreover, AI is also blamed for other disruptive features in the legal profession such as the problems of complexity, the worrisome increasing autonomy of AI systems over time, the problem of opacity in decision making of AI systems, and the technological vulnerability of AI systems because they are highly dependent on collected data, which may be insufficient, inaccurate, or biased. 189 Moreover, the fact that AI systems are exposed to cyber-security attacks or breaches is a major challenge to the development of legal AI. 190 There is also a legitimate concern regarding who is going to be responsible for the mistakes of AI tools that are just a piece of machinery or a program, the developer or the users. Therefore, not a few numbers of scholars suggest that the crucial need to regulate and hold someone accountable shall be treated by the law. 191 On the other hand, there are also other major noticed constraints to integrate AI into the legal profession such as technical constraints, the problem of the complexity of legal reasoning, the lack of adequate market for legal AI (economic constraints), and the significantly slow culture of legal practice (cultural constraints) to integrate with AI. 192 Moreover, studies show the likely negative impacts of legal AI technologies on the legal profession in general such as high rate of unemployment 193 , insecurities related to data privacy, ethics, and dishonest use of data, and the unwanted creation of a super-intelligent AI, which is also called the 'Singularity problem'. 194 On the other hand, there is a justifiable argument that judges should not delegate judgments or a specific administrative task to an AI assistant and judges need to stay in full control. 195 Moreover, AI is in principle deemed incapable of adequately engaging in legal (analogical) reasoning or evaluative judgments, which is considered a serious challenge to legal AI in the long term. 196 Moreover, the fact that many judgments involve an element of discretion, which is not the case for computer programs that operate based on the logic of input and output exacerbated the problem of integration of AI with the legal profession. 197 Finally, there is also a convincing argument that AI is not immune from the bias and prejudice of its creators as a result; it cannot be always trusted to be fair and neutral. 198

Conclusion
In this essay, an effort was made to enumerate and discuss the future impacts of AI on legal research in the legal profession. As some scholars try to portray, the law is neither rocket science nor entirely repugnant of technology. Hence, legal research in particular and legal practice, in general, is amenable to and influenced by AI both positively and negatively.
Moreover, it is evident from the study that the positive impacts of AI are far greater than its negative externalities, which are usually temporary and related to the disruptive effects of technology on the legal profession.
It should also be emphasized that legal research, which includes multifaceted activities is a core lawyering skill and an integral part of legal practice. All types of legal professionals (judges, lawyers, legislators, and academicians) must undertake legal research in due course of delivering various types of legal services and the quality of their research determines the quality of the services they provide to clients. As a result, when one tries to assess the impact of AI on legal research, s/he is also implicitly assessing such an impact on the entirety of legal practice to which the research is an integral part.
Only some decades ago legal research was an activity that can only be done by lawyers in a physical library. At present, due to advances in "Weak AI", many of the activities that constitute legal research are being done by AI tools with minimal human support, which resulted in monumental efficiency (in time, energy, resources) in the underway of legal research and legal grunt work.
At present, there are up to 5000 legal tech startups throughout the world who are automating some type of legal work, which is a good reminder for tomorrow's lawyers that they will need to familiarize themselves with how to research the law using such AI tools in addition to 197  In the future, with the advent of Strong AI, which has a massive computational and analytical capacity of a vast amount of data and brute force of processing, the impact of AI on legal research will be far greater than mere automation (pre-programmed decision making). With such a leap in computational capacity and advances in algorithmic reasoning, AI tools are expected to develop the capability to deliver efficient legal services by autonomously undertaking legal research that is destined to sort out legal problems that will require human empathy, judgment, and creativity and thereby satisfy client expectations.