Written by Thomas Lee (Associate Author) | Mentored by Ong Chin Ngee | Reviewed by Rakesh Kirpalani
LawTech.Asia is proud to have commenced the third run of its popular Associate Author (2020) Programme. The aim of the Associate Authorship Programme is to develop the knowledge and exposure of student writers in the domains of law and technology, while providing them with mentorship from LawTech.Asia’s writers and tailored guidance from a respected industry mentor.
In partnership with the National University of Singapore’s alt+law and Singapore Management University’s Legal Innovation and Technology Club, five students were selected as Associate Authors. This piece, written by Thomas Lee and reviewed by industry reviewer Rakesh Kirpalani (Drew & Napier and DrewTech), marks the second thought piece in this series. It scans the landscape of lawyers and technology, and sets out steps that lawyers should take to meet a future technologically-driven paradigm.
Introduction
In the United States, a study by McKinsey estimates that 23% of work done by lawyers in the United States can be automated by existing technology given the emergence of new artificial intelligence software that can scan legal documents, streamline communications, and find relevant case work for lawyers.[1]
Consequently, this paints a worrying picture within the Singaporean context – that technology may disruptively challenge and redefine the role of lawyers in the foreseeable future. In his 2019 address at the Garden Gala Dinner of the Council of ASEAN Chief Justices, the Honourable Chief Justice Sundaresh Menon warned that there is simply no reason for the legal community to think that the transformative potential of technology would not play out in altering the legal scene at a frightening speed – just as it has done in other industries.[2] In the same vein, this article briefly discusses how new technologies have shaped the legal scene in Singapore, and what implications this will bring to the legal profession in general. To this end, it is argued that lawyers will need to diversify their skillsets to become tech-literate in order to optimise workflows in a law firm to ultimately improve the quality and delivery of legal services. It is also argued that lawyers should hone “soft” skills that cannot be replaced by technology but will be increasingly useful in practice.
Technology that can be used as an assistive tool
Lawyers are increasingly able to rely on a host of digitised programs to perform a variety of legal tasks. This new wave of novel technologies pervading the legal sphere has largely been kickstarted by legal start-ups. One of the most prominent faces in this arena is Luminance – a company created in 2015 by leading mathematicians from the University of Cambridge. Luminance’s platform utilises machine learning algorithms to identify patterns and warning signs in legal documents in areas such as Mergers & Acquisitions (“M&A”), property and compliance.[3] Notably, a single Luminance platform draws upon a database of over 5,000 uploaded contracts, unlocking capabilities for M&A lawyers such as (1) tracking regional activity and governing law that demonstrates the geographic spread of the contracts and alerting users to surprise areas that may pose risks, (2) detecting patterns in any language or jurisdiction, (3) comparing different types of clauses in the contract against other similar clauses in contracts within its database to locate what is different from the norm.[4] This changes the M&A scene greatly by significantly improving productivity: lawyers who once had to rigorously review every contract in the entire data room without losing the context of the wider transaction can now more easily do so by relying on this platform to zoom in on contracts that are more likely to present issues that require their attention.
In a separate area of law, such as family law, the rules of the game have also shifted with the introduction of a simulator that predicts the division of assets in a divorce proceeding. This simulator started as a project from Lex Quanta, which predicts the outcome of the division of matrimonial assets by computing factors such as the total amount of assets involved and how each party contributed to the relationship, through an algorithm.[5] This benefits members of the public, who may use a simplified version of the simulator to get a rough sense of the problem and see if they should hire a lawyer. Today, this simulator has since been tested with more than 10 family law firms, and its founders are further looking to expand the simulator’s capabilities to other cases involving personal injuries sustained in traffic accidents, and commercial cases such as contract and intellectual property disputes.[6] Following this trend, it is perhaps safe to predict that such technologies would soon permeate every area of law in the foreseeable future. In turn, on-the-job training to become tech-literate would easily be a necessary skillset for all current and future legal practitioners.
Limitations of legal AI – the need for human emotion and empathy in technology
Nonetheless, all artificial intelligence (“AI”) systems used in legal work to-date suffers from one key limitation – they do not address the personal, social and emotional aspects of the law. For instance, in matrimonial matters, parties may remain undecided in initiating divorce proceedings. While an outcome simulator like Lex Quanta’s may help litigants gauge a probable outcome should they proceed with a divorce, they still require a human touch: a family lawyer who can guide them through all their available options (legal or non-legal) to ascertain the option most beneficial to their interests. This is particularly critical in sensitive and emotionally-complex issues such as custody or care and control, as outcome simulators (or any existing AI system for that matter) cannot give a qualitative answer to what will be in the best interests of a child. It would there be unrealistic and not beneficial to believe that legal AI systems are fully capable of replacing lawyers just yet. By extension, this presents an opportunity for lawyers to grow an emphasis on human judgement and empathy to deliver the most beneficial and sustainable results to clients.
Further, there is a potential for unintended bias to occur in any AI solution. This happens due to the innate nature of biases within data used to train an AI system, or biases inherent in the people and processes involved in developing or deploying such AI systems, which are then reflected in the AI solutions that they develop. For instance, a recent study that analysed the Correctional Offender Management Profiling for Alternative Sanctions (“COMPAS”) software package used by some US courts to predict the likelihood of recidivism in criminal defendants found that COMPAS was as accurate as a small crowd of nonexperts.[7] Understandably, such a predictive tool has been met by sentencing courts with scepticism and close scrutiny, in part due to reliability issues given how recidivism models may not be based on factors that are objectively reliable. One is reminded of the adage “garbage in, garbage out”, which highlights how any algorithm is only as effective a predictive model can be depending on the factors it is being programmed to assess. Oftentimes, judges must apply soft skills when speaking to accused persons, and by gauging their demeanour, adjust sentences accordingly to ensure that justice is not only done but is also seen to be done in the eyes of the public.[8]
Another example of machine bias in the legal field concerns online legal case databases. It was found that engineers who designed these search algorithms for case databases such as Casetext, Fastcase, Google Scholar, Lexis Advance, and Westlaw had inherent biases on what would be considered a relevant case – biases that were then reflected in the search algorithms surfacing cases to users.[9] While there may be ways to mitigate or address such biases – such as taking steps to ensure data quality and to train the AI solution on representative data sets, or (for the lawyer) to balance out some bias by researching in more than one online database – the concern remains the bias that goes unnoticed and percolates through the legal system.
Overall, this demonstrates another need for human lawyers – to review the predictions or decisions raised by an AI system, and to remain the final accountable actor in any decision that will impact a client. To be able to do so, however, lawyers must become more literate in these technologies, and be cognisant of the benefits and risks that arise with the use of technology solutions. For instance, in using AI systems, lawyers may have to be diligent in understanding the quality and representativeness of the training data that such predictive solutions rely on, before relying on these technologies.
Unavoidable technology within the courtrooms
Finally, it is worth noting that where new technology is adopted in the courtrooms, lawyers (especially litigators) would need to follow suit and embrace such changes. In Singapore, our judicial system has demonstrated a receptive approach to adopting new processes to facilitate wider access to justice. In July 2017, the Small Claims Tribunals (“SCT”) launched the Community Justice and Tribunals System (“CJTS”), an online case filing and management system that allows parties access to Community Justice Tribunal’s services online. Since the launch of the CJTS for the SCT claims until the end of February 2019, 1725 small claims have undergone e-Negotiation, and 602 cases reached amicable settlement.[10] Looking at these statistics, the CJTS has arguably been a successful – a point perhaps reflected in its adoption elsewhere, such as the Community Disputes Resolution Tribunals and the Employment Claims Tribunals. Further, during the COVID-19 pandemic, the Singapore Courts allowed hearings to be conducted through Zoom, allowing lawyers and parties to attend court remotely and safely. This latter example perhaps speaks best to the growing indispensable nature of technology in the delivery of legal services and justice.
Conclusion
Back in the 1960s, a Stanford University computer scientist Roy Amara observed that people “tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run” – an observation now known as “Amara’s Law”.[11]Indeed, the fast-moving technological changes that surround the legal industry today as compared to a decade ago bears a good reminder that lawyers should learn to be tech-literate, and at least understand how to use these emerging technologies as a complement to their lawyering skills. In the same vein, even for more well-established firms, putting a brand above the door does not necessarily prepare lawyers to deliver efficiently in the future, and top firms in Singapore must also keep up with such changes to continually offer competitive rates in the market since clients ultimately value cost against efficiency. Thankfully, to help law practices sharpen their competitive edge, the Government has incentive schemes for law practices to apply for, such as the Productivity Solutions Grant to encourage firms to continue their digitisation and productivity upgrading efforts.[12] By leveraging on these grants, even smaller firms can remain competitive in this industry.
[1] (n.d.). Retrieved from https://public.tableau.com/profile/mckinsey.analytics#!/vizhome/AutomationandUSjobs/Technicalpotentialforautomation
[2] See the Garden Gala Dinner address delivered by The Honourable the Chief Justice Sundaresh Menon (23rd November 2019). Accessible at https://cacj-ajp.org/closing-address-gala-dinner-23-nov
[3] https://www.artificiallawyer.com/2019/12/09/al-product-review-luminance-part-one/.
[4] M&A Due Diligence Case Study (Rajah & Tann Asia). (n.d.). Retrieved January 17, 2021, from https://www.luminance.com/files/case-studies/rajah_tann.pdf
[5] Koh, F. (2018, January 07). NUS law and economics student, along with three peers, creates case outcome simulator. Retrieved from https://www.straitstimes.com/singapore/nus-law-and-economics-student-along-with-three-peers-creates-case-outcome-simulator
[6] Koh, F. (2018, January 07). Legal startup’s simulator predicts who gets what in divorce. Retrieved January 17, 2021, from https://www.sgsme.sg/news/startups/legal-startups-simulator-predicts-who-gets-what-divorce
[7] Dressel, J., & Farid, H. (2018). The accuracy, fairness, and limits of predicting recidivism. Science Advances, 4(1), p.2.
[8] Re Shankar Alan s/o Anant Kulkarni [2007] 1 SLR(R) 85 at [103].
[9] Mart, S. N. (2016). The Algorithm as a Human Artifact: Implications for Legal {Re}Search. SSRN Electronic Journal. doi:10.2139/ssrn.2859720.
[10] See the State Courts Workplan 2019 delivered by Justice See Kee Oon, Presiding Judge of the State Courts, “State Courts” 2020 and Beyond” (8 March 2019).
[11] Definition of Amara’s law. (n.d.). Retrieved January 17, 2021, from https://www.pcmag.com/encyclopedia/term/amaras-law
[12] Sim, C. (2018, December 26). Government Grants. Retrieved from https://www.lpi.lawsociety.org.sg/government-grants/
This piece was written as part of LawTech.Asia’s Associate Authorship Programme.
Featured image credit: eDepoze