Opinion | Inclusion Of AI In Judicial System Offers A Range Of Practical Benefits, But Caution Is Essential
The use of artificial intelligence should be seen as an ‘assistance’ tool, not as a replacement for human adjudication. Over-reliance on automated processes may conflict with the principles of discretion and fairness
Being a pleader with one of the oldest professions of the world, when you look at the society and the myriad of interpersonal relationships and complexities, the role of a lawyer becomes extremely important. As our society has grown, the legal profession has become structured. As a result, we are facing an epidemic-like condition of pendency of cases.
As of 2024, India faces a severe backlog in its judiciary, with over 44 million pending cases, including 82,989 in the Supreme Court alone, according to the National Judicial Data Grid (NJDG). Despite a disposal rate of nearly 94.92 per cent, the backlog remains significant, with cases continuing to accumulate due to disruptions caused by the Covid-19 pandemic. The Supreme Court is currently managing a large number of cases, including 1,130 before a three-judge bench, 274 before a five-judge bench, and others before even larger benches. To address this, a Special Lok Adalat was organised in July-August 2024, successfully resolving 1,100 out of 2,200 listed cases.
related stories
This backlog not only delays justice for individuals but also hampers economic progress, creating a growing need for innovation. However, there is rising optimism surrounding the rapid development of artificial intelligence (AI). AI offers tools that could revolutionise the way courts operate, potentially reducing delays and increasing access to justice for millions.
AI’s Potential to tackle pendency
AI refers to computer systems designed to perform tasks that typically require human intelligence, such as learning, problem-solving, and decision-making. Unlike human intelligence, AI relies on algorithms and vast datasets to identify patterns and make predictions. The Supreme Court has taken initial steps toward AI integration. SUVAS (Supreme Court Vidhik Anuvaad Software), launched in 2019, employs machine learning to translate Supreme Court judgments into nine regional languages. This aims to make justice more accessible by breaking down language barriers.
Another notable tool is SUPACE (Supreme Court Portal for Assistance in Courts Efficiency), which was launched to provide digital infrastructure to support the digitisation of court processes. SUPACE helps reduce the time spent on routine tasks by reading case files, extracting relevant facts and arguments, and presenting them in a concise manner for judges and researchers. It also assists in drafting routine documents, effectively improving efficiency and streamlining judicial processes. Additionally, SUPACE incorporates machine learning features that can mimic human behaviour, similar to the algorithms used by platforms like YouTube, where videos similar to those previously watched are suggested. By automating administrative tasks, SUPACE frees up time for judges to focus on the speedy resolution of cases.
Several Indian High Courts are experimenting with AI applications. In Jaswinder Singh v. State of Punjab (2023), the Punjab and Haryana High Court made an unprecedented move by seeking input from ChatGPT to gain a broader understanding of global views on granting bail in cases involving cruelty. The Delhi High Court is implementing AI Saransh, a tool that generates summaries of legal pleadings to help judges quickly grasp the key issues in a case, further expediting the judicial process.
Globally, AI is also making strides in judicial systems. Argentina’s Prometea automates repetitive bureaucratic tasks, extracts relevant information, and even prioritises cases based on urgency. Brazil’s VICTOR system uses Natural Language Processing (NLP) to analyse documents, significantly reducing the workload of the Supreme Court. In Estonia, AI judges resolve small claims disputes, while Canada uses AI to handle minor property and motor vehicle claims.
Challenges
While AI has started to be integrated into the Indian judiciary, bias is one of the biggest concerns arising from the datasets used to train the algorithms. If the data reflects existing societal prejudices, the AI system may perpetuate or worsen these biases, leading to unfair outcomes. A study by the University of Pennsylvania found evidence of “in-group bias" within Indian law enforcement agencies, particularly against individuals from Scheduled Castes and Scheduled Tribes. Introducing AI systems trained on such biased data could exacerbate these inequalities, further disadvantaging already vulnerable communities. For example, the COMPAS system, used in the US for sentencing recommendations, has been found to be biased against African Americans, raising concerns about its potential to perpetuate racial disparities.
Another key challenge is the “black box" nature of many AI systems. The lack of transparency in AI systems’ decision-making processes can undermine public trust and make it difficult to ensure accountability. In Australia, the Supreme Court expressed similar concerns about the potential for AI-based predictive models to perpetuate existing biases against indigenous communities in Director of Public Prosecutions for Western Australia v. Mangolamara.
The potential for AI “hallucinations" also poses a significant risk. Generative AI systems can fabricate information, potentially leading to inaccurate legal advice. In Roberto Mata v. Avianca Inc. (2023), a Manhattan federal judge fined a lawyer $5,000 for submitting fictitious legal research generated by ChatGPT with the intent to mislead the court. Additionally, in the Delhi High Court case involving Christian Louboutin, Justice Pratibha M. Singh rejected the use of ChatGPT-generated responses, citing concerns about potential inaccuracies, fictional case laws, and imaginative data generated by AI chatbots.
Suggestions
The Supreme Court’s experiments with AI tools like SUVAS are a positive step toward reducing translation errors in critical documents like FIRs or witness statements, which often result in prolonged delays in the disposal of cases. AI can also play a crucial role in better case management and prioritisation. Its use in random case allocation can eliminate manual intervention, ensuring that the process remains unbiased and fair. By analysing data, AI can predict case outcomes and identify cases that are more likely to be resolved quickly or require urgent attention, allowing courts to prioritise those cases that need immediate action. Additionally, integrating AI into routine and straightforward matters, such as civil violations, drunk driving, and cheque bounce cases, could lead to faster resolutions and a reduction in overall backlog.
While the inclusion of AI in the judicial system offers a range of practical benefits, caution is essential. The use of AI should be seen as an “assistance" tool, not as a replacement for human adjudication. Over-reliance on automated processes may conflict with the principles of discretion and fairness. There are legitimate concerns about biased datasets and the potential risks to individual rights. Enhancing the efficiency of the judicial system should not come at the cost of undermining equity and the human aspect of justice. Therefore, the inclusion of AI should be approached with careful consideration and caution.
The author is a Senior Supreme Court Advocate and former Additional Solicitor General of India. Views expressed in the above piece are personal and solely those of the author. They do not necessarily reflect News18’s views.
- Location :
- First Published: