*Aditya Gatlewar


The robotic judge that was once only seen in science fiction films has finally become a reality. The COVID-19 pandemic underlined a greater need for innovation for resolving disputes as humans vacated our court rooms. But with machines potentially more innovative and considerably faster than the human brain, allowing them to operate as judges may pose a threat to the legal system. This article  examines the risks and benefits of such technology to evaluate whether it is a threat or a boon.


The right to “Access to Justice” has been integral to the common law and democratic society. However, today’s technological revolution has brought us to the essential question of “Which Access To Justice”? Is it Techno-Logical or Human-Logical? Artificial Intelligence (“AI”) has rhetorically become a part of our lives. Today, democratic society has been shaped so that Artificial Intelligence and Technology have become indispensable. For example, previously, an opinion had to be expressed through limited newsletters, journals, etc., but today, everyone with access to a variety of internet platforms can express an opinion, and readers have the option of leaving ratings and comments. This online platform uses artificial intelligence to display content that interests viewers.[1] AI is already playing a critical role in the media, which is the fourth pillar of democracy. Montesquieu’s vision of separation of power with branches of government into Legislative, Executive and Judiciary was to ensure that no one branch is more powerful than another.[2] However, with the emergence of Artificial Intelligence, all three branches, especially the judiciary, are in the step of replacement. If not governed by proper laws, AI may take an absolute position in all three branches in the near future. The burden of the pendency of cases is the most crucial challenge to the Judiciary. To reduce the burden, various parts of the world are on the verge of using Artificial Intelligence in the judicial system. The article is to explore the risk involved in this replacement and finds out whether such a replacement is a threat or a boon.

Prevalence of AI in the Judicial System

The Covid-19 pandemic has brought significant speed to the development of the industry. Many countries like China and Estonia were ready to rapidly develop Internet-based AI-Judges to avoid the social spread of Covid-19 in the courts. In December 2021, China became the world’s first country to make an AI-equipped judge, who is said to give 97% correct decisions after listening to oral arguments. These judges can hear theft, credit card fraud, and dangerous driving matters.[3] The capacity to analyse billions of megabytes of data in the system certainly gives it an edge. However, from 2017 China has handled millions of specific cases like trade disputes, copyright infringements and e-commerce liability claims by the robot–judge. Estonia was also using a robotic–judge for small disputes. The Estonian Ministry of Justice in March 2019 asked “Ott Velsberg” (Government Chief Data Officer of Estonia) and his team to design an “AI or Robotic Judge” that could adjudicate upon small claims disputes up to €7000.[4]

The roles of mediators and lawyers had already proven to be functional and beneficial before the arrival of AI-judges. The “Robotic Mediator” had its first success in February 2019. It settled a dispute in which a personal counselling trainer claimed outstanding fees on behalf of a client. Mr Graham Ross, Mediator and ODR expert, proposed to his clients to use the Canadian Dispute Resolution programme “Smartsettle ONE” after phone mediation failed.[5] The “DoNotPay” app, made by Joshua Browder, is the world’s first robot lawyer and has  spread widely across the UK and US. With 150,000 subscribers, it helps them to write letters on issues like insurance claims, applications for tourist visas, complaint letters, cancelling gym memberships etc. This robot lawyer has also won an award from the American Bar Association for increasing legal access.[6] India, the largest democracy in the world, has the most ‘pending’ court cases and therefore AI has received strong support from Hon’ble law minister “Kiren Rijiju.” Former CJI[7] Hon’ble “SA Bobde” stated that “AI can help in reducing the backlog of cases and sustainable delivery system.” However, India has not yet reached authorised a Robot- lawyer or Judge. But, the country has welcomed a mix of human and AI portal known as “Supreme Court Portal for Assistance in Courts Efficiency” or “SUPACE”, whose job is mainly focused upon collection and analysis of data.[8]

The AI-Lawyer or AI-Mediator do not expose the judicial system and the parties to too much risk, as there is always a human judge to check and verify. Conversely, a judge being the final decider, the AI-Judge is the biggest risk to the judicial system as the role of a judge is complex and sensitive. It is necessary to set a boundary between technology and humans. The newly developed technology should only be considered to be assistance and not become a substitute for a human judge. To understand this, it is important to draw a comparison between the conceptual abilities of technology and humans.

Artificial Intelligence (“AI”) V. Human Intelligence (“HI”)

As per the study of neuroscience, “The “Mind” is energy, and it generates energy through thinking, feeling, and choosing. It is our aliveness, without which, the physical brain and body would be useless. That means we are our mind, and mind-in-action is how we generate energy in the “Brain”, and “Brain” merely responds to these unique experiences”. [9] This is how Human Intelligence works; it is our mind that makes our brains responsive. Technology today can create brains and use them for judging cases. Nevertheless, the question is, “whether this technology would make any dissimilarity as compared to human judges for not being alive or not consisting of mind”. To answer this question, we need to understand how the Robot-judge works.

The Robert-Judge functions mainly on two disciplines of AI; Machine Learning (“ML”) and Natural Language Processing (“NLP”). ML can be defined as a machine capable of imitating intelligent human behaviour. The best examples are chatbots, predictive text, language translations and showing feed, which interest viewers.[10] And, NLP can be defined as “an automatic manipulation of natural language”, like speech and text, by software.[11] It gives the ability to a computer program to understand spoken and written human language.[12] Prominent examples of this are Alexa, Google Assistant, Voice text messaging etc. This system works by combining various data with fast iterative processing and intelligent algorithms. Looking at the entire structure of work, it can be inferred that Artificial Intelligence works by a straight rule bonded within the borders of what is expressed through algorithms. However, such structure for the judicial system can create the following problems:

  1. Black Box Problem: AI-Judge working upon inputs and outputs based upon algorithms gives no visibility on the process and working of AI. We cannot know what this machine does and how it does it. This can create a problem for the AI-Judge in interpreting the text written in a statute or judgement.
  2. AI tools use a “deep-machine” learning system: The AI’s opacity can challenge human abilities to explain how they work and ensure fair outputs. AI learns its own translations, it may be accurate, but such a function lacks transparency.[13]
  3. AI tools are powerful in narrow settings: The fast pace can increase difficulties for traditional policy tools where it may prioritise business concerns over societal values.[14]
  4. May violate due process rights: If there has been a historic judgement in the United States that was discriminatory against black people. A machine will reference such cases and create cognitive biases in the decision.[15]

Even though AI is very effective and can surely reduce the burden upon the judges and courts, keeping in mind the above problems, an AI-Judge can never be a substitute for Human Judges. This technology should only be considered a complementary service to human decision-making authority. In discussing AI v. HI, Justice Surya Kant said, “If e-technology would be allowed to overpower the judicial field without any ‘Lakshman Rekha’, are we marching towards a stage of engaging robots in place of judicial officers?”[16] The question is valid. However, the answer to this question is in the story itself, where “Justice Surya Kant” has taken the word “Lakshman Rekha”.

The incident is found in the Ramayana of Indian Mythology, which hints at what may happen if AI is entirely replaced with HI. During their Vanvaas (forest exile), Sita, Rama’s wife, noticed a deer in the jungle. Rama left by telling his brother Lakshmana to watch after his wife after she asked him to fetch it for her. However, Rama did not appear, so Sita asked Lakshmana to locate him. Lakshmana objected, but Sita persuaded him to go and get Rama. Lakshmana had to survive owing to Sita’s pressing order, so he drew a line outside the hut and asked Sita not to step outside. Suddenly, Ravana appeared in the form of a Sadhu, or saint, asking for food. Sita offered him, but he said he would only accept such alimony if she stepped out of the line and fed him. Sita crossed the line, and Ravana abducted her.

The line is commonly known as “Lakshman-Rekha” and was intelligent enough to not allow outsiders into the house. However, such intelligence cannot understand the manipulation made by Ravana to abduct Sita. If Lakshmana was there, the same incident would never have happened. Hence, AI can never replace HI.

Conclusion: Threat or Boon

As discussed above, any human when working uses his mind first and then responds to the brain. In the study by “DL Chen”, it was observed that external factors like the tone of words used in the first three minutes of a hearing, the incidence of birthdays, the outcomes of sporting events, and even the time of day a hearing or the defendant’s name affect the outcome of cases, according to studies conducted in the United States, France, Israel, the United Kingdom, and Chile.[17] Bringing AI into the judiciary system will eliminate these external factors. However, as discussed in the event above, “Lakshman-Rekha” cannot act similarly to “Lakshmana” himself. Similarly, a Machine cannot act as a judge.

Following Montesquieu’s objective of separation of power, all three forms of government should only be taking assistance and support from AI and it should not be a substitute over these three bodies. Under the separation of power theory, a judge cannot be a legislator or executive. However, AI is performing in all three divisions. There is no doubt that AI is developing rapidly and can grow faster than a human brain; under such circumstances making AI part of government is overall risky and is a threat. However, its lifelong promise of ‘assistance only’ and humans not showing laxity in their entrusted work will make it a boon to the judiciary and humanity.

[1] MIKE KAPUT, what is artificial intelligence for social media, MARKETING ARTIFICIAL INTELLIGENCE INSTITUTE (Feb. 18, 2022, 0.32 AM), https://www.marketingaiinstitute.com/blog/what-is-artificial-intelligence-for-social-media#:~:text=How%20is%20AI%20used%20in,uses%20AI%20to%20identify%20visuals.

[2] LEGAL INFORMATION INSTITUTE, https://www.law.cornell.edu/wex/separation_of_powers_0 (last visited Feb. 17, 2022).

[3] HINDUSTANNEWSHUB, https://hindustannewshub.com/world-news/china-made-the-worlds-first-artificial-intelligence-equipped-judge-gives-97-percent-of-the-decisions-right/ (last visited Feb. 17, 2022)

[4] WIRED, https://www.wired.com/story/can-ai-be-fair-judge-court-estonia-thinks-so/ (last visited Feb. 19, 2022)

[5] NICK HILBORNE, Robot mediator settles first ever court case, LEGALFUTURES (Feb. 20, 2022, 09.47 AM), https://www.legalfutures.co.uk/latest-news/robot-mediator-settles-first-ever-court-case.

[6] PADRAIG BELTON, would you let a robot lawyer defend you?, BBC (Feb. 21, 2022, 02:14 AM), https://www.bbc.com/news/business-58158820

[7] Chief Justice of India.

[8] SREEJANI BHATTACHARYA, Can AI Revolutionise India’s Judicial System?, ANALYTICSINDIAMAG (Feb. 21, 2022, 10: 41 AM), https://analyticsindiamag.com/can-ai-revolutionise-indias-judicial-system/

[9] CAROLINE LEAF, How Are The Mind & The Brain Different? A Neuroscientist Explains, MBG, (Feb. 26, 2022, 05:51 PM), https://www.mindbodygreen.com/articles/difference-between-mind-and-brain-neuroscientist

[10] SARA BROWN, Machine learning, explained, (Feb. 26, 2022, 07:25 PM), https://mitsloan.mit.edu/ideas-made-to-matter/machine-learning-explained

[11] JASON BROWNLEE, What Is Natural Language Processing?, Machine Learning Mastery, (Feb. 26, 2022, 07:33 PM), https://machinelearningmastery.com/natural-language-processing/

[12] TECHOPEDIA, https://www.techopedia.com/definition/653/natural-language-processing-nlp (last visited Feb. 26, 2022)

[13] JEFF WARD, 10 Things Judges Should Know About AI, (Feb. 27, 2022, 04:31 PM), https://judicature.duke.edu/articles/10-things-judges-should-know-about-ai/

[14] Id. 12

[15] ATANU BISWAS, A robot in a judge’s chair, (Feb. 27, 2022, 04:31 PM), https://www.thestatesman.com/opinion/robot-judges-chair-1503031697.html

[16] Id. 14

[17] SANDEEP BHUPATIRAJU, DANEIL L. CHEN, SHAREEN JOSHI, The Process of Machine Learning for the Courts of India, (Feb. 27, 2022, 06:54 PM), https://nlsir.com/the-process-of-machine-learning-for-the-courts-of-india/#_ftnref6