An Overview on Legal Data Science
Legal data science is the nexus of law and data analytics. At its core, it involves applying statistical, computational, and analytical methods to legal data with the intent of solving a legal problem or otherwise enhancing the understanding of a legal issue. This is the technical definition of legal data science; however, the field of legal data science is more than just the analytical process. The practitioners often need a deep understanding of the legal practice as well as the means by which data analysis is performed. Thus, legal data scientists, as they are commonly called, are not only lawyers but also computer scientists of some sort.
One thing which is abundantly clear is that legal data science is on the rise. We recommend that practitioners who wish to explore the intersection of legal and data go no further than legal technology. Legal technology is essentially just the application of technology to the delivery of legal services , which encompasses a wide range of topics but is nonetheless tailored very much to the day to day practice of law. In fact, even if we were being more abundantly accurate, this entire blog is legal technology. The reality is that legal technologists deal with the myriad of legal data issues on a daily basis, whether it be deciding how to best communicate with a client, or organizing a law firm’s knowledge bank. It is perhaps in this sense that legal data science is the parent of legal technology; while every legal technologist can act as a legal data scientist, the reverse does not apply. But that does not mean that many other fields of practitioners will not continue to increasingly dabble in the art and science of legal data science.
Use Cases in Legal Practice
Data science is being used in various applications throughout the legal industry. From contract review to litigation support, data scientists are creating software to make a range of processes faster and more efficient.
In one form of application, Contract Review, data scientists have developed tools such as LawGeex and LegalSifter that use Artificial Intelligence (AI) to review contracts. In the case of LawGeex, the AI is trained on 1200 contracts from 30 industries. A human lawyer reviews the contracts along with the AI, but the AI will give its suggestions faster and effectively ‘recommends’ edits for the human to consider. Researchers at Luminance, a Software-as-a-Service solution for document review, tracked 100 documents and showed that Pangea3 (now part of E&Y) lawyer reduced the amount of time spent reviewing documents from 500 hours to 140 hours. Time is money for legal practitioners, and the importance of efficiency cannot be overstated. However, Pangea3 lawyers still needed the help of 20 AI lawyers to analyze the 100 contracts.
Another example of a contractual application is LawGeex which uses Natural Language Processing (NLP) to analyze contracts and provide recommendations for edits. LawGeex is used for reviewing contracts in house and by law firms. For example, a corporate lawyer will upload a contact into LawGeex and it will ‘read’ the contract and highlight inconsistencies, errors or problem areas. The lawyer will then have the ease of looking at the potential issues instead of having to sift through the whole document searching for them. Of course, for ethical and procedural reasons, The Lawyer is still needed for review, but the aim of these innovations is to make contract review easier and more efficient.
A third application is Decision Analytics, which uses Predictive Analytics to help lawyers and their clients make decisions on how to proceed in a case. For example, in a litigation matter, a client may want to know what the likely outcomes of a case are. Rather than going to a lawyer who may have a imperfect recollection of past judgments, lawyers who specialize in Decision Analytics may have the most relevant data at their fingertips. One innovative regulatory firm, Knowyourcustomer, uses a Range-of-Probable Outcomes (RPO) estimate to calculate the probability that a client is compliant with Anti-Money Laundering legislation.
A final example is visualizations that can be used to tell a story, such as the following interactive chart created using LinkedIn data about lawyers:
Benefits for Legal Solicitors
Legal professionals who embrace data science into their work will find that it brings efficiencies and informs their decisions. They will be able to leverage predictive analytics, natural language processing, and other AI to help them better represent their clients and serve their businesses. Case Law Analytics, for example, helps attorneys support their arguments with historic cases and rulings. Thompson Reuters HighQ uses AI to assist lawyers in identifying risks to protect internal interests and those of specific clients. Docket Alarm provides attorneys insight on filings including impending deadlines resulting in substantial savings in e-discovery. Salinno Markets Inc. offers lawyers a number of AI tools including NOVA, a client acquisition tool that matches clients with the right attorney. Legal Evolution, a leading legal marketing and growth platform, helps law firms to increase client acquisition rates, build client loyalty, and more effectively price legal work. Luminance uses machine learning to read thousands of pages of documents helping lawyers spot critical changes in contracts and other documents. Gazelle.ai monitors company datasets to help professional services industry members like lawyers to target clients with business growth potential. These are just a few of the legal AI companies growing in the industry.
Challenges and Ethical Implications
In the context of legal data science, challenges can arise from a range of sources. As a discipline at the intersection of law and technology, legal data science inevitably involves sophisticated technological tools, and the emergence of machine learning technologies in particular. One key challenge here relates to data privacy: organizations need to be careful about processing personal data in ways that may give rise to breaches of privacy law, both in terms of the data used for training and testing algorithms, and any personal information that is generated as part of the testing and training process. A wide range of techniques may mitigate this challenge.
Another challenge relates to the need for these technologies themselves to be better understood. As the field of legal data science continues to evolve, it will be important to ensure that the underlying algorithms used for processing data are thoroughly tested for reliability, and for any risks of bias or other problems that may disproportionately impact certain groups. It is also essential to ensure that machine learning technologies, in particular, are explainable. Related to this is the need to ensure that lawyers understand how the implications of the use of these technologies in their work .
A related, and more global, challenge lies in the level of integration of these technologies and legal data science into the practice of law. For example, lawyers will need to periodically update and maintain their underlying infrastructure if they are using data-driven technologies, but the general integration of these systems into legal practice is just beginning and presents challenges to innovators, users, and regulators alike.
Specialized skills in data science itself are not always easy to come by in the legal sector. Change is coming, but slowly. While many schools of law, around the world, have already integrated courses on legal technology, legal information management, artificial intelligence, and related fields into their curricula, it is clear that there is much more work to do in legal education. The Canadian Bar Association has even proposed an overarching Competency Framework for Lawyers, which outlines a series of competencies that "reflect both contemporary practice and the future of practice, encompassing technology-related skills and knowledge as they might give rise to professional responsibility obligations."
Examples and Case Studies
The application of legal data science is rapidly making an impact throughout the legal world. A few current examples are: E-discovery: Leading e-discovery vendors are using predictive coding (called TAR 1.0 and TAR 2.0) by teaching their computers to make predictions of whether a document is relevant or not. Through training the computer on samples of documents previously reviewed, the computer becomes "smarter" and generates its own predictions for the remaining documents under review. Because computers can be trained at a faster and larger volume, e-discovery is made more efficient for both the law firm and the client. Forecasting: Several jurisdictions now provide access to predictive analytics to forecast future litigation and its probable outcomes. An example is Lex Machina, which includes "litigation analytics" that is based on machine learning and natural language processing of thousands of cases to assist clients and attorneys with making decisions concerning legal strategy and risk management. Portfolio Analysis: Premonition is an example of a data analytics technology that provides a litigation intelligence platform, developed to identify relevant "lawyers and firms". As described on its website, Premonition "scans millions of cases every day and extracts court dates; it can instantly tell you how lawyers and judges perform in the cases that matter (or will matter) for you. Predictive Practice Analytics: BakerHostetler, Landslide Innovations, and Thomson Reuters have come together to build BaketRack, a predictive practice analytics to anticipate the types of legal issues and client demands their attorneys will face. Leveraging artificial intelligence and facts mined from 100 years of BakerHostetler files backed by Thomson Reuters corpus of law, this predictive app is being woven into the fabric of the practice, enabling BakerHostetler to predict outcomes of motions, when to settle, and what retaining clients are watching. A competitive edge: Many law firms have been using legal data science technologies for several years to weed out the high volume of false data and deliver the critical information and insights attorneys need for case analysis faster than competitors. The technology helps firms identify legal research content with the best coverage for their practice and substitute artificial intelligence for the human factor for faster, more transparent, repeatable, and scalable results.
Prospects for Legal Data Science
The future of legal data science is likely to involve further integration of machine learning and AI technologies, providing a powerful new tool for legal professionals. As more firms and in-house departments adopt these technologies, the competition between them may increase, prompting firms to find ways to differentiate and provide unique value to their clients.
Lawyers may begin using legal data science internally as well, streamlining and automating routine tasks and making data-driven decisions about how best to allocate their time and resources. On the other hand , legal consumers may use new technologies to provide better access to affordable legal information and services.
Legal data science will also have an impact on the courses that current and future lawyers must take, and on the information that they must track after graduation to enhance their professional competency. Some legal data scientists even recommend that law schools teach traditional legal concepts alongside these new skills, reflecting the changing legal landscape. Finally, whether or not lawyers are able to use legal data scientists, legal data analysis will soon require lawyers to track data across mega projects—projects with 1 million or more records—and be able to point to the reasons that were critical for winning or losing the case.