The ever evolving and hasty advancement of artificial intelligence(“AI”) is permeating various sectors at a very fast pace, the judiciary being no exception to it. However, the incorporation of AI algorithm models into the judicial system has gained a lot of attention and has been a matter of intense debate and interest worldwide, including in India. Lately, what has been uncovered is that Indian High Courts are employing AI models like ChatGPT to aid in case management and legal research. Foremost, the Manipur High Court and the Punjab and Haryana High Court came into the limelight by taking a pioneering step in this regard, and setting up a potential precedent for the possible application of AI in the legal domain. This leap by the Indian High Courts also beseeches for a thorough analysis of leveraging AI into legal adjudication.
The Allure of AI: Early Implementation
On 23rd May 2024, while delivering a judgment, the Manipur High Court undertook the assistance of ChatGPT in conducting research before passing the order concerning the Village Development Force (“VDF”).
The said petition in the high court was filed by Md. Zakir Hussain who was disengaged from the service (VDF) in 2021 without being given an opportunity to be heard and without conducting any formal inquiry into the allegation of dereliction of duty. Justice A Guneshwar Sharma, heading a single-judge bench inquired about the procedures for the disengagement of VDF personnel. Further, it held that the disengagement orders by Superintendent of Police Thoubal were in violation of the principle of natural justice, quoting the judgment of the Nasim Banu Case. The court accordingly set aside the order of removal and restored the petitioner to the services.
Justice Guneshwar Sharma, after requesting the government counsel for information regarding disengagement procedures of VDF personnel and not getting the same, was then compelled to research around important information pertaining to VDF with the assistance of ChatGPT. Moreover, the court acquired knowledge of the establishment, functioning, roles and responsibilities of VDF in Manipur, which was initiated to strengthen local security and assist local police, specifically in rural areas.
This was not the first instance of the use of AI to aid in delivering justice and in March 2023, the Punjab & Haryana High Court emerged as a pioneer in taking assistance of ChatGPT in legal research. Justice Anoop Chitra employed ChatGPT to get insight into bail jurisprudence involving ‘cruelty’ in homicide cases and received knowledge indicating the possible denial of bail in such circumstances while deciding on the case involving the accused Jaswinder Singh, who was alleged to have assaulted an individual, which led to his death.
The Broader Picture: Indian Courts on AI
Delhi High Court’s Skepticism
Later in the same year, the High Court of Delhi chastised lawyers from taking ChatGPT or any other AI model in use. In August 2023, Justice Pratibha Singh stated the observation by the Delhi High Court that lawyers cannot use ChatGPT to provide reasoning on ‘legal or factual matters in a court of law’. She expressed caution over the potential use of ChatGPT and AI Models by legal practitioners to render “incorrect information, imaginative data or fictional laws”, a trend that has been witnessed in other countries as well.
Supreme Court’s Balanced Approach
In contrast to the Delhi High Court, the Supreme Court of India has attained a very measured and balanced approach to the incorporation of AI in the judicial system with time. Successive Chief Justices from S.A Bobde to D.Y Chandrachud all have often referred to AI as a “tremendous asset”, nevertheless striking a balance between vigilant adoption and intrigued execution. Supreme Court has even constituted the Artificial Intelligence Committee chaired by Justice L Nageshwara Rao. In a significant move, the erstwhile Chief Justice of India, Justice S.A Bobde, launched an Artificial Intelligence (AI) based portal ‘SUPACE’, which stands for the Supreme Court Portal for Assistance in Court’s Efficiency for the Judiciary with the modus operandi of assisting judges in the collection of relevant law and facts to the case. From time to time, the Supreme Court came up with different AI aided models for different purposes like live transcription, judicial translations, E-Courts projects, etc., upholding the paramountcy of AI in the legal framework.
Ethical Concerns and Potential Risks
Privacy Risks and Data Protection
The integration of AI along with its potential to revolutionise the judiciary, also brings in critical concerns regarding the privacy and data protection of an individual, which cannot be overlooked. AI systems in courts involve the collection huge amounts of data, encompassing the personal data of the parties involved, evidence, financial records, medical records, and other confidential information. Though the Digital Personal Data Protection Act, 2023 (“DPDPA”) enforces strict restrictions on the whole process of collection, usage and storage of personal data, the intricate, automated, and opaque data processing operations entailed by AI are inconsistent with the foundational principle of DPDPA. which poses a significant risk in safeguarding the personal data of any individual involved in the judicial process
Bias and Accountability
The potential existence of bias is a major concern with AI in the judiciary. As AI models rely on historical data to produce the desired result, the result may be prejudiced toward certain issues that exist in the legal domain. If these prejudices are not rectified, AI may prolong and even worsen the injustice in the court’s decision. This makes the dependence of the judicial system on AI models for legal assistance worrisome.
Secondly, accountability is also a major concern. In this context the question that arises is that if AI makes an erroneous decision, who must be held accountable for that? AI models are not legal practitioners and scholars. They do not take into account ethical standards and legal prerequisites while providing legal advice or desired results. The core of the problem centers around the transparency of AI algorithms, often described as ‘Black Boxes’. The more and more we rely on AI for decision-making especially in sensitive domains like the judiciary, the question of trust becomes a pressing concern. It is because it is still unclear how AI reaches its desired outcome. The employment of advanced algorithmic deep neural networks by AI models, by their very nature, remains opaque and largely hidden from human comprehension. This lack of transparency is often regarded as the ‘black box’ snag in AI. These problems pose significant challenges to the usage of AI bots by Judges, which can’t be ignored.
The Human Element in Justice
For centuries, justice has been a human-centric endeavour and has been centred around individual rights, interests and a deep understanding of the complexity of human nature. The erstwhile approach formed the system like human rights, constitutional rights and international law in a way to be governed by human activity. A human judge is influenced by factors such as available information, cultural and social background, etc, which are major prerequisites in ensuring personal integrity in doing justice. On the other hand, the AI model are fed data containing historical societal biases and other discriminatory practices during its training which would eventually produce flawed results, leading to potential injustice. Therefore, the onset of AI in the judicial system raises concerns over external interference and potential injustice.
Conclusion: Charting a Balanced Integration
It is undeniable that the hefty number of pending cases overburdens the Indian Judiciary. At this point, delivering justice at a fast pace with no gaps for disparities is the need of an hour and taking the assistance of AI to make the process swift and less time-consuming is a groundbreaking move. However, the onus in the last must rest upon the judges for their final verdict, taking into consideration broad implications, ethical concerns and human conscience. Also, legal professionals must be trained with a proper methodology to effectively use and analyse AI outputs.
Furthermore, in a democratic nation like India, where judicial independence is a prominent concept, leveraging AI in the legal process must be critically analysed before the integration of the same. As in the current scenario, we lack guidelines and regulatory frameworks to administer the use of AI by our legal fraternity. Therefore, in order to address this concern, the enactment of certain sets of rules and regulations is of paramount significance. This should include the minimisation of collection and retention of personal data without hampering the functioning of AI and leveraging to only take into use the data that is required for producing the desired outcome. Data anonymisation and pseudonymisation are the techniques that may help in fostering the same. Furthermore, these regulations must also mandate the AI models to adhere to relevant privacy regulations and standards, such as the General Data Protection Regulation (GDPR). A correspondent integration of AI with all regulatory measures taken supplements the legal process working under human oversight and does not overtake them, quenching the need of the hour.
The post is written by Aryan Panwar, 2nd year BALLB Student from RMLNLU.
