Sarbeswar Mishra* & Taskin Akhtar**
Introduction
“The future of artificial intelligence is not to replace man with a machine, but to enhance man’s capabilities.”
Artificial intelligence (“AI”) is not a story of fiction; rather, it is a legal reality. AI has become an integral part of the justice delivery system, a reality that is shaping legal education, courtroom arguments, legal aid platforms, and legal databases throughout the world. Google’s Sundar Pichai has quoted AI’s purpose as to expand human capabilities, rather than replacing them. AI serves a plausible, if partial, solution to the pending cases and constrained resources of the Indian legal system. Yet such a shift raises the concern of fault, prejudice and ethics. This article examines AI’s transformative possibilities in law alongside the risks it poses, seeking to determine whether AI poses a threat to the system or serves towards a more advanced egalitarian legal system.
Turbocharging Legal Research with AI
The AI-driven legal-research platforms, such as Casetext, LexisNexis AI, and Manupatra, allow practitioners to do so research within seconds in respect to statutes, case laws and precedents by means of machine learning and processing of natural language. These tools reduce the gap between newcomers and advanced practitioners, reducing hours of research and improving accuracy. Most importantly, these tools particularly benefit Indian law firms’ demanding efforts, with the rise of caseloads and limited resources. Although the empirical picture is subtler than what is commonly assumed, the efficiency gains are abundant. In a report, Thomson Reuters’ Future of Professionals Report, which surveyed across fifty countries, including 2,200 legal professionals, concluded that respondents expect AI to annually save roughly 240 hours, worth an estimated $32 billion in the United States industry-wide.
In May 2025, it was found out in a survey by the Manupatra Academy that more than 50% of Indian Legal professionals, including those in criminal practice, already use AI on a daily basis, and around 73.7 per cent had used generative AI. However, the primary barrier to adoption is data privacy tension. According to the National Judicial Data Grid (NJDG) of 2024, India faces a backlog of cases exceeding around 50 million. In such a position, AI helps in reducing human error and speeding up research, but over-dependency on AI poses a risk, such as failing to notice jurisdiction-specific or customary issues that can only be solved by human reasoning. A generic model may disturb India’s numerous customary laws. One should maintain the equilibrium between AI’s speed and the irreplaceable human counsel and situation-specific judgment.
Navigating Jurisdictional and Regulatory Divergence
In the legal profession, an intermittent drawback of AI is that law is not limited to one system, but rather numerous systems; an interface trained mainly on a particular jurisdiction’s case law and drafting conventions will not automatically transfer to another. In Common-law systems such as India, the U.K., the U.S., and Australia, they rely on precedent-based reasoning; on the other hand, civil-law systems codify rules extensively and subordinate precedent. In such a scenario, a prototype that is well-articulated in one jurisdiction can misappropriate the other. India amalgamates with a plural legal landscape, statutory law over common law, separate personal law and customary law regimes governing marriage, succession and religious practice, that can be presently navigated without human verification by a few AI trained globally.
The divergence in regulatory frameworks is equally pronounced. The European Union’s AI Act take up a binding, risk-tiered approach classifying systems as unacceptable, high, limited or minimal risk. The United States has no analogous federal statute; its professional responsibility is addressed through ethics opinions. In the American Bar Association’s Formal Opinion 512 (July 2024), it was held that lawyers who use generative AI should execute their existing responsibility of competence, confidentiality and supervision. The Bar Council of England and Wales updated its non-binding guidance in November 2025, stressing verification of AI output and protection of confidential data.
India has no binding cross-sector AI statute. NITI Aayog’s National Strategy for Artificial Intelligence (2018) and its Approach Document for Responsible AI (2021) remain advisory rather than enforceable, built around safety, equality, inclusivity, transparency and accountability. In this vacuum, individual courts and bar bodies have issued mutually inconsistent rules. The Kerala High Court’s Policy Regarding Use of Artificial Intelligence Tools in District Judiciary (July 2025). The Bombay Bar Association’s Guidelines on AI use (July 2025), the first bar-association-level guidance in the country, and a Punjab and Haryana High Court circular (April 2026) barring judicial officers from using tools such as ChatGPT, Gemini and Copilot for research or judgment-writing. The Supreme Court’s AI Committee has since released Draft Regulations for the Use of Artificial Intelligence in Courts, 2026 for public consultation, proposing that AI may assist but never adjudicate, that a human must remain in the loop, and that its use in court documents be disclosed. The difficulty compounds with cross-border data. An Indian cloud-based AI legal tool must ensure the rules of the Digital Personal Data Protection Act, 2023 (“DPDPA”), alongside the Advocates Act, 1961, the Information Technology Act, 2000, and, where relevant, the EU’s General Data Protection Regulation. A tool built for one jurisdiction’s evidentiary standards cannot be assumed to be compliant in another.
Streamlining Legal Operations Through Automation
AI tools, such as ContractPodAi, eBrevia, and Kira Systems, help in reducing the strenuous work of drafting and reviewing. It reads a multitude of clauses of a document within minutes and identifies risks, which may take a junior associate hours to flag. It mechanises contract review and monitors compliance with fair accuracy. In India’s cost-conscious legal market, this efficiency allows firms to handle advanced caseloads without an increase in staff proportionately. In 2026, the Wolters Kluwer Future Ready Lawyer survey found that more than 90% of legal professionals nowadays use at least one AI tool, mentioning efficiency and reduction of cost as the most important advantages. On the other hand, the commonly faced difficulties in further adoption are ethical concerns and data privacy.
Adoption of AI divides the legal market into two parts; however, this division is less understood and anticipated by the professionals. The initial aspect focused on by everyone is the efficiency: automation of document review and preliminary research reduces firms’ preliminary and recruitment costs. However, on the other hand, the Wolters Kluwer survey itself cautions that it reduces firms’ demand for associates and at the same time retains traditional legal skills, thus not affecting the firm but affecting the job market. AI-enabled compliance platforms allow the corporate teams to monitor regulatory obligations on their own, rather than the traditional approach of consulting outside counsel for lower-value work. However, the strings of logical arguments of economics were countered by a 2025 study by the National Bureau of Economic Research. It observed that generative AI chatbots had no statistically significant effect on the working hours or wages across professions, including legal professions. Thus, the real demand shifts towards strategic and verification work rather than wholly substituting lawyers with AI.
Prompt engineering becomes a minimum requirement to fully realise the potential of these systems. However, many Indian professionals lack the skill even at its basic level, thus limiting the tools’ usefulness even in accomplished hands. It becomes important to realise that a small error in a prompt, or an algorithmic mismatch, can compromise deals, drafted contracts and above all, the security of the job of the user; a small technical mistake would probably cost millions of rupees as well as jobs.
These concerns are not merely hypothetical or anecdotal. India’s paralegal workforce occupies an uncertain position. The parent statute, the Advocates Act, 1961, which governs the profession, does not formally recognise “paralegal” as a distinct category. This leaves the job opportunities of such professionals uncertain in jurisdictions with formal AI-adoption guidance. The Bar Council of India (“BCI”) has publicly taken the position that, as AI tools are not legally recognised in India under the Advocates Act, only the advocate using the tool can be held responsible for erroneous AI-generated content. However, any benefit arising out of such AI use is enjoyed by the firms. In the authors’ view, this asymmetry, where accountability is on advocates, the job at stake is of paralegals; however, the efficiency gains captured by firms create a structural issue in the policies. This issue needs to be addressed by a BCI-issued framework rather than the present patchwork of advisory notes and circulars. AI should be used as a tool to provide efficiency to working professionals rather than putting their jobs at stake; in application, AI should provide job security rather than job threats.
Enhancing Access to Justice for the Masses
Justice for all, square and fair, is the promise made by AI. Technology enables democratised justice where everyone, irrespective of income or location, has equal access to AI tools. It creates a scope to reduce historical prejudice caused by structural barriers. Services such as DoNotPay, Rocket Lawyer, and India-based platforms such as MyAdvo and LawRato ease access to justice by providing access to legal information, document templates or vetted lawyers directly to litigants and freshly practising advocates. Such apps bridge the gap between access to justice and economic need.
AI shows its true efficiency by providing legal acumen to non-lawyer citizens. For instance, the DoNotPay chatbot has enabled people to challenge their parking fines and pursue small claims without a lawyer. In the transition of the paradigm, NITI Aayog’s 2021 Approach Document identifies safety, equality, inclusivity and transparency as guiding values for AI usage. AI-assisted tools serve as a good associate with schemes, making schemes more accessible. For instance, the National Legal Services Authority (“NALSA”) paralegal volunteer programme, assisted with AI tools, can help address queries away from over-stretched legal-aid lawyers, increasing the efficiency of schemes. AI tools widen the scope of justice for rural litigants by bridging the gap of linguistic differences between the legal language and indigenous tongue. However, it is important to develop an efficient system which actually bridges the linguistic requirement. Lessons should be taken from past manoeuvres. For instance, the Supreme Court’s own translation tool, SUVAS, had to be built indigenously because generic AI translation performed poorly across India’s twenty-two scheduled languages.
It is important as well to realise the forte of AI use and not to use it recklessly. AI performs well in systematic, clerical and laborious tasks, such as creating Excels and List of Dates, while it’s reliability reduces with increasing complexity of tasks, such as a plaint filing and legal research. In such instances, the user needs to be careful about how he uses the tools for his benefit. In Jaswinder Singh v. State of Punjab, the Court consulted ChatGPT only for a broader, comparative view of bail jurisprudence; the court clarified that the AI response was not intended to guide or substitute the court’s own reasoning. As in Christian Louboutin, the Indian courts have treated AI output as, at most, a preliminary research aid, it never allowed it as evidence or a basis for adjudication. The Indian legal sphere rightfully focuses more upon human-verified checkpoints, thus on a positive track towards a balanced and fair ecosystem by realising the efficiency as well as the challenges which AI posses.
Addressing AI’s Challenges and Solutions in Law
Other than its fascinating advantages, AI carries along with it the risk of prejudice. The recommendation of AI gets sourced from earlier court decisions on similar points of law; however, it misses the aspect that the law being upheld is case-specific. Such embedded data results in biased outcomes, reducing legal reasoning to mechanical statutory interpretation. Reliance on AI for precedents narrows the scope of applying the legal mind and narrows the scope for judicial reasoning that is rational and case-specific. Algorithm-suggested responses ignore the social dynamics and demographic needs of India as a diverse nation. Vehement reliance on AI makes prompt engineering a necessary skill set for lawyers in the near future. AI systems raise concern as it remains “black boxes”, leaving users clueless about the legal principles, jurisprudence and rationale behind a response. In cases when AI fails in a high-stakes matter, it raises serious questions of accountability, resulting in a separate suit to determine who was responsible, whether the developer of AI, the firm that took the case or the advocate who did the AI search.
The concerns of AI hallucination is no longer a hypothetical scenario. An order passed by the Bengaluru bench of the Income Tax Appellate Tribunal on December, 2024 had to be recalled upon the discovery that reliance was placed on Supreme Court and Madras High Court citations that never existed. The same issue of fabricated precedents was faced by the Bombay High Court in October 2025. This concern is not indigenous to India; in the American case of Mata v. Avianca, Inc.,where attorneys were sanctioned under Rule 11 of the Federal Rules of Civil Procedure for filing a brief citing six wholly fictitious, ChatGPT-generated judicial opinions. Such incidents concur alongside the Delhi High Court’s caution in Christian Louboutin that AI falls in the “grey area” of reliability. Considering the current position of Indian law, where the liability falls on the advocate who signs and files the document, exposes the individual making him liable to systemic technological limitations, which are beyond the individual’s control. Such a position of liability makes the individual reluctant to explore the scope of efficiency generated by the use of AI.
AI also raises concerns of data privacy. Privacy finds its place among the fundamental rights guaranteed under Article 21, post Justice K.S. Puttaswamy (Retd.) v. Union of India, judgement. This being the premise, direct constraints and questions flows against AI, to how the AI systems handles client data, case records or personal information? Whether such operation is lawful, and whether it adheres to the data-fiduciary obligations that apply once the DPDPA, 2023 Rules take full effect on 13 May 2027. However, there are answers to these raising concerns, such as NITI Aayog’s advisory Approach Document for Responsible AI (2021), which tries to navigate AI systems toward fairness, transparency and accountability.
It lays down a guide to answer the concerns of safety, reliability, equality, inclusivity and non-discrimination by AI usage. One of the solutions to address bias is through audits, such as the framework proposed by IBM, a useful model for the profession. The basis of legal arguments should not be mere faith in AI-generated results; rather, the professional should strictly cross-check the responses with the statutes, case laws and case-specific requirements.s. The concerns of job security may be dealt with through training in prompt engineering and algorithmic literacy, opening new roles in AI oversight. However, the Indian paradigm requires a robust statutory framework, beyond such advisory principles. The forthcoming Draft Regulations for the Use of Artificial Intelligence in Courts, 2026, address these needs and would serve as a foundation for future legislation. However, we need to consider that any policy would require continuous inputs from technologists, lawyers and policymakers, to sustain it; otherwise, the boon turns into a fatal risk.
Conclusion: AI in the Courtroom, Vision or Risk?
AI has brought tremendous change in the field, especially the legal atmosphere of India, where pendency of cases and uneven access to justice remain the major issues. However, other than being efficient, AI carries real challenges of bias, transparency and accountability. The survival of an AI-driven legal system depends upon a balanced approach between efficiency and principles of fairness and justice. Contemporary developments in the form of the Supreme Court’s 2026 draft regulations and various court-level guidelines illustrate the need for a balanced approach. Human control should remain the focus of legislatures, to make the algorithms function for, rather than against, the principles of justice. AI should be developed with the moral scaffolding, while preserving lawyers’ capacity to exercise their own judgement and intellect, thus retaining the human touch over the discourse. The legal profession should shift its focus from fascination with efficiency to defining the roles permitted to AI in serving justice and accountability. Focus should be on the way of transforming the legal system and the aftermath of such transformation. AI is part of the future of the legal domain, but the direction of that change should lie in the hands of humans. The article is a plea to the profession to balance technological advancement with the enduring pursuit of justice.
* The author is a third-year law student at the National University of Study and Research in Law, Ranchi. The author may be contacted at sarbeswar.mishra@nusrlranchi.ac.in.
** The author is a third-year law student at the National University of Study and Research in Law, Ranchi. The author may be contacted at taskin.akhtar@nusrlranchi.ac.in.
This blog reflects the personal views of the author and does not necessarily represent the views of The Policy Chronicle.