March 18, 2019
March 21, 2019


Should we use Artificial intelligence in a Lawyer’s Office?
Artificial Intelligence scares me. As I write this, the immediate thought in my mind is, -‘What about my job?’. India has several unemployed Lawyers, with less than 10% making a living wage[1]. Despite holding a professional degree, Lawyers are so heavily unemployed. We can take the pessimistic view that the Legal Profession will stop requiring human capital when AI can be employed. Such a conclusion could be true but the legal profession itself also comes with a variety of notable problems which Artificial intelligence could fix.
The Problems
Lawyers are generally perceived to be slow in adapting to change. For example, in India Stamp duty payments for contracts, it must be done manually[2]. Share certificates in India are required to be held onto physically when there is a share purchase[3]. The loss of the certificate requires a determination by a Registrar as to who has title to the shares. Having a digital share certificate would solve the problem. Law Firms in rely heavily on books for legal solutions, despite solutions being easily accessible online.
With the context above, the skepticism of Lawyers in employing AI in their offices is validated. Where many lawyers are struggling even to move beyond books, using a machine to decide legal fates seems far cry. Yet, the usage of AI can fix some pressing law firm issues.
AI: The Future?
Mental Health
Working hours, conditions and client deadlines are scientifically one of the biggest contributors to poor mental health amongst employees[4]. Employees are prone to quit as lawyers have one of the highest rates of attrition amongst professions[5]. The flowing consequence of work culture -> poor mental health -> quitting isn’t an individualistic phenomenon, but a systematic problem. Artificial intelligence can address this problem.
Large amounts of documentation are required for M&A transactions, requiring lawyers to conduct a due diligence process. The process takes long and requires immense time commitment with tight deadlines. More importantly, no human intelligence is required for this work. A study on Artificial Intelligence by LawGeex shows it to be faster than humans at reviewing 5 NDAs – in just over 25 seconds with 97% accuracy. The best human took 51 minutes, while the average was 92. The best human was only at an 85% accuracy[6]. The study involved some of the brightest young lawyers.
Smarter Lawyers
Resarch to find the most basic cases can takes hours, which is frustrating. ROSS AI is a tool which has been employed by some premier law firms, with users reporting increased productivity and quality of work from lawyers[7]. The tool finds the most appropriate cases, or statutes that can be applied in a particular legal situation. It places reliance on an individual’s intellectual enterprise rather than so called ‘hard work’, creating smarter Lawyers. The lawyers learn to represent research and present better solutions to their clients.
The goal of afirm is to make profit. AI can contribute to this. It generates more work in less time, allowing firms to increase their productivity. In the LawGeex study, there is a clear gap in the time taken to process legal review – while the AI was under 26 seconds, the average human was at around 92 minutes. The legal work doesn’t end when the AI finishes its processing, humans will review the work and present it to the client in an accessible manner. AI nuances the work of a lawyer increases the productivity of the firm. Previously, a lawyer could process approximately 20 NDAs a day (7 hours of work + breaks), given that the average lawyer took 92 minutes. With AI tools, this time could be cut to a quarter, perhaps lesser. The Lawyer will only need to conduct a basic review. In addition, some time will be taken to prepare for a client meeting. The time that will be taken for a re-check and preparation for a client meeting aren’t even taken into account in the 92 minute estimate. Removing the time taken for actual review will likely lead to heavily increased productivity. Dispensing with work quickly will result in greater clientele.
Access and Affordability
Lawyers charge high hourly legal fees, in exchange for heavy workload. The pricing leads to poor access to justice. Clients can expect to win a case solely due to hiring a better lawyer. Lawyers are also fair in understanding that the work timings are such that costs should be high. Artificial intelligence engines like WestLaw Edge are exceptionally effective tools, a legal search that gives direct answers to complex legal questions[8]. The AI calculates the probability of success of the argument as well. This allows the lawyer to present the most relevant information in court in exchange for shorter working hours. Increasing clients with lower pricing will still increase profits.
Maybe it’s just better?
It is important to understand that AI is just better than us. Not only is AI faster, it is also more accurate – uses a much larger knowledge base, can review documents that a keen human eye may miss and learns faster than any human can. It replaces ‘menial’ legal tasks, like reviewing contracts. Reviewing a contract for favorable clauses and unfavorable clauses can be more easily effectuated by a system like LawGeex. Lawyers can do smarter research, argumentation and presentation and client negotiations, instead of drudgeries and banal work. However, this isn’t just a sales pitch for AI. There are notable issues in implementing it, and perhaps we can find solutions to these problems.
Should we trust AI?
Picture this: you type a bunch of words on the screen, and a machine decides your fate. AI requires humans to begin trusting and believing in the machine. A recent study at The University College of London highlighted that an AI judge was able to predict with 79% accuracy, the judgments of the European Court of Human Rights[9]. A similar result was found with a US Supreme Court study at 83% accuracy[10]. However, AI does not have 100% accuracy, this is partly due to hindsight bias and partly a result of limited emotional and social awareness. The information processing method of an AI bot is algorithmic and largely ignores the capability for political, moral or social thought. For example, schemes that are employee beneficial, but come with an employer cost are likely to be ignored by an AI bot. AI is likely to view it as irrational, despite building a better work environment. The question then becomes, how much can we trust it?
An absurd AI use is ‘Wevorce’[11], which is an application where couples wishing to get divorced can list their situation and choose their favorable outcomes while the AI can explain the likelihood of success of those outcomes. It seems almost jarring to think of using AI in dictating things like child custody, which is an emotional matter. In fact, the process of divorce itself hinges on an emotional appeal. AI also has the ‘accountability’ problem, wherein unlike a human conducting due diligence, a robot cannot be held responsible when committing an error in its process. Granted, the chances of errors are a lot lower, but nevertheless the possibility of the same is fair warning.
Future in AI
When AI comes into the fore, it is possible that a machine takes over our lives. However, there is no reason why AI cannot be curbed. It is important to realize the value of AI and find smart and unique ways to govern its usage such that it can be used effectively. Instead of letting AI take over, AI should act as a facilitator to create smarter lawyers and be able to attack problems like legal costs and pendency. The Government could regulate it’s use. For example, it can only be employed as a research tool and not an adjudicatory tool. Further, its use can be banned when the issue is to do with family law, or a question of serious social dynamics. Another issue could be the use of AI as competitive advantage. The Government can also use its power to normalize the legal field and allow AI to become a more affordable for small law firms. With AI we can fix some of the most notable problems to customers and lawyers and its value shouldn’t be underestimated or dismissed on account of skepticism.


[1] Dubey RC, “Are Lawyers Obsolete?” (Business TodayDecember 7, 2018) <https://www.businesstoday.in/opinion/columns/are-lawyers-obsolete/story/298367.html> accessed February 9, 2019
[2] Indian Stamps Act, 1899. ss. 10.
[3] Companies Share Capital and Debenture Rules, 2014. ss. 6.
[4] Flaherty S, “How Clients Contribute to Mental Health Problems at Law Firms” (The Legal IntelligencerOctober 24, 2018) <https://www.law.com/2018/10/24/how-clients-contribute-to-mental-health-problems-at-law-firms/> accessed February 9, 2019
[5] Christin L, “Confronting Lawyer Turnover in Law Firms” (Attorney at WorkNovember 13, 2018) <https://www.attorneyatwork.com/confronting-lawyer-turnover-in-law-firms/> accessed February 9, 2019
[6] Leary K, “The Verdict Is in: AI Outperforms Human Lawyers in Reviewing Legal Documents” (FuturismFebruary 27, 2018) <https://futurism.com/ai-contracts-lawyers-lawgeex> accessed February 9, 2019
[7] “How a U of T Legal Research Startup Is Making Lawyers More Productive: Wired Magazine” (University of Toronto NewsNovember 20, 2017) <https://www.utoronto.ca/news/how-u-t-legal-research-startup-making-lawyers-more-productive-wired-magazine> accessed February 9, 2019
[8] Ambrogi B, “Move Over Westlaw – Meet the Next-Generation Westlaw Edge, With Advanced AI and Analytics” (LawSitesAugust 7, 2018) <https://www.lawsitesblog.com/2018/07/move-westlaw-meet-next-generation-westlaw-edge-advanced-ai-analytics.html> accessed February 9, 2019
[9] Mascarenhas H, “New Artificial Intelligence ‘Judge’ Can Predict the Outcome of Human Rights Trials” (International Business Times UKOctober 24, 2016) <https://www.ibtimes.co.uk/new-artificial-intelligence-judge-can-predict-outcome-human-rights-trials-1587938> accessed February 9, 2019
[10] Hutson M, “Artificial Intelligence Prevails at Predicting Supreme Court Decisions” (Science | AAASJuly 26, 2017) <https://www.sciencemag.org/news/2017/05/artificial-intelligence-prevails-predicting-supreme-court-decisions?r3f_986=https://www.google.com/> accessed February 9, 2019
[11] Rivera C, “A Start-up Offering Divorces Online” (CNBCApril 6, 2016) <https://www.cnbc.com/2016/04/06/a-start-up-offering-divorces-online.html> accessed February 9, 2019


Samvit Ganesh is a 4th Year Law Student from Jindal Global Law School and his hometown is Bangalore, India. The author takes a keen interest in the subjects of Intellectual Property Law, Environmental Law and Company Laws. The author is passionately engaged in social work while engaging with multiple NGOs in India.




In Content Picture Credit: Bigger Law Firm

Leave a Reply

Your email address will not be published. Required fields are marked *