By — Vridhi Parakh
ABSTRACT:
In the 21st century, it is impossible to live without consuming food, water and Artificial Intelligence. Its impact is so domineering that it’s concluded that in the near future, AI will take your jobs. The professions that practice with justice in the country thought that they were safe from this attempt at domination. However, given certain developments by enthusiastic tech-ventures, the artificial red light has threatened to take over the law. The advent of paralegal AI bots like jhana.ai has come with the power to rationalise Indian common law, strengthen the stare decisis system and democratise access to legal research. This article argues that while AI-powered legal research tools help overcome the burden of the justice system, they risk enacting an algorithmic colonisation of the law, subtly determining the precedent for the Court, all without regulatory scrutiny and democratic accountability.
The column lays down a four-pronged argument for the same: reinforcement of legal hierarchies at the expense of pluralist and customary legal traditions, reproduction of historical biases embedded in training data, advent of an access-divide of tools to the already privileged and a vacuum in legal education that allows these platforms to enter law schools through peer networks rather than through a critical pedagogy. There is still a long way to go before AI can be used in a way that it doesn’t eradicate the social factor of law.
INTRODUCTION:
The Indian Legal System has long operated inside a paradox- it sustains a multi-layered, pluralist judicial architecture while simultaneously dealing with a backlog of over 51.5 million pending cases. The Supreme Court alone disposes of tens of thousands of matters each year. High Courts across the country produce a further torrent of decisions, many of which never reach the databases that junior advocates can afford, and some of which conflict openly with each other. Into this landscape of productive chaos, AI legal research tools have arrived as the black-cloaked saviour: find the right case, synthesise the applicable ratio, and surface citations in seconds.
The optimistic case for an AI Paralegal rests on the age-old question of the efficiency of the Indian Judicial System. Article 141 of the Indian Constitution binds Supreme Court decisions on all courts throughout India. While it sets in stone an ascending ladder of judicial hierarchy in theory, it dissolves whenever a practitioner cannot locate the dominant precedent, when conflicting decisions from coordinate High Court benches remain unresolved, or every time a district court advocate in a non-tier 1 city lacks a subscription to SCC Online or Manupatra. The bridge between the doctrine of stare decisis and its practical implementation is structural, and it widens the access-gap basis geography and income.
AI Paralegals that can successfully retrieve factually and legally consonant precedents can, in principle, close this gap. Former CJI D.Y.Chandrachud described AI in legal research as a game-changer that empowers legal professionals with unmatched efficiency and accuracy. A system in which cases wait years for resolution cannot afford to dismiss tools that reduce research time and ease workloads. This gives rise to the problem of who decides what the algorithm finds. The boundaries of what counts as the law or important precedent are drawn by a venture-backed startup operating under commercial incentives that have nothing to do with justice.
THE PRIVATISATION OF PRECEDENT
The appropriation of social life by data relations as a new form of extraction goes beyond historical resource colonialism to reshape the very categories through which reality is perceived. In the legal context, when a private platform determines which judgments surface at the top of a search and which sink to page forty, it performs a privatisation of precedents. The paralegal, who may have historical biases embedded in its algorithm, has the power of constitutional impact without its ‘checks and balances’. AI models founded on Global North data and values frequently yield biased outcomes that exacerbate inequalities, prioritising the legal categories of dominant traditions over those of marginalised ones. The legal traditions that AI fails to surface simply become less visible, less citable, and, over time, less legally salient. This leads us to the transparency problem.
This article uses the term ‘algorithmic colonisation’ to describe the process by which computational systems, designed and controlled by private technological actors, begin to shape legal knowledge by determining which authorities, interpretations, and categories become discoverable within legal reasoning. Jhana.ai does not publicly document what its relevance algorithm prioritises, how its dataset handles unreported decisions or what its hallucination rates look like at scale. It is incompatible with judicial use. A judiciary whose research is pre-structured by a proprietary algorithm that no party can examine, challenge, or appeal is a judiciary whose historical independence has been partially surrendered to a private actor.
ALGORITHMIC BIAS AND THE REPRODUCTION OF LEGAL INEQUALITIES
Among the poorest 20 per cent of households, only 2.7 per cent have access to a computer and 8.9 percent to internet facilities. The communities most exposed to legal vulnerability (tribal forest rights claimants, domestic workers, migrant labourers, marginalised women navigating personal law, among others) are precisely the communities least likely to benefit from a subscription-based, bandwidth-intensive, knowledge-heavy AI research platform. LLM systems trained on historical legal data inherit the biases of that history. Indian legal history chronicles caste hierarchy, gendered disproportion, and colonial administrative categories that shaped both the substance and the procedure of law. A model trained on reported decisions from a system that historically underserved Dalit and Adivasi litigants will reproduce the reasoning patterns of that underservice, not correct them. Research on algorithmic bias and data colonialism confirms that algorithmic systems are effective only for populations for which training data exists. Where training data is sparse, or where it reflects historical patterns of exclusion, the algorithm’s outputs replicate and amplify those exclusions.
LEGAL EDUCATION AND THE FUTURE OF PROFESSIONALISM
Perhaps the most underexamined dimension of the AI paralegal revolution is its entry into Indian legal education. Generative AI has not entered Indian law schools through curriculum reform, faculty consensus, or institutional policy. It has entered through peer networks: one student recommending an AI tool to another, sharing prompts, and exam season producing a quiet proliferation of AI-assisted drafts. In India, the Bar Council of India proposed introducing modules on law and technology, but implementation remains pending. In the interim, students use LLM tools to draft memos, produce research summaries, and analyse documents without any structured instruction on what these tools can and cannot do, whose legal epistemologies they encode, or what professional responsibility attaches to their outputs. The danger of dependence this creates is one of phantom precedents and mass unprecedented AI-powered law. A generation of lawyers trained to accept AI outputs as research, rather than as a starting point requiring rigorous verification, will reproduce the artificial assumptions of those tools in their professional practice, which will more or less end in embarrassment.
In Mata v. Avianca, attorneys were sanctioned for submitting a brief that cited six entirely fabricated judicial decisions generated by ChatGPT. The court found that the attorneys had abandoned their professional responsibilities by failing to verify AI-generated outputs through any conventional legal research tool. More troublingly, when opposing counsel challenged the citations, the attorneys asked ChatGPT to confirm its own output, and the model obligingly confirmed that the hallucinated cases were real. Similarly, India is experiencing a rise of reliance on algorithmically-generated, fabricated judgments in the Courts.
Law schools must respond to this environment not merely by teaching students how to use AI tools, but by teaching them to interrogate those tools. Critical pedagogy in this context means asking foundational questions: Who built this tool? Whose legal traditions informed its training data? What does it exclude, and why? How do I verify its outputs independently? Law schools must adapt curricula to include AI literacy, ethics, and hands-on critical learning, ensuring that students develop the capacity to evaluate and responsibly use AI in legal contexts rather than simply consume its outputs. Anything less is an abdication of the pedagogical function that the legal profession requires.
CONCLUSION
The AI paralegal platforms arriving in Indian legal practice are not villains. They address a genuine crisis. An Indian judiciary with 51.5 million pending cases, a research infrastructure that prices out district court practitioners, and a professional culture that demands speed without always rewarding accuracy, does need better tools. But efficiency is not the only value that Indian law must serve. Justice requires that the law be accessible to the subaltern, not just the subscribed and that the lawyers and judges who use these tools do so with eyes open to their assumptions, their exclusions, and their failure modes.
None of this is possible without regulatory action. India requires, at a minimum, three urgent interventions. First, the BCI must issue professional guidance specifying that AI-generated research requires independent verification before citation. Second, the draft Supreme Court’s guidelines for transparency and auditability for AI usage in Courts must be implemented. Third, Indian law schools must embed critical AI literacy into their core curricula.
The challenge is not simply to make AI more powerful, but to ensure that the power it acquires remains answerable to the Constitutional principles upon which the Indian legal system itself is built. A legal system committed to constitutional values cannot allow the determination of legal relevance, precedent, or access to justice to be governed solely by opaque private systems. The future of AI in law must therefore be shaped by democratic oversight, professional responsibility, and institutional safeguards that ensure technology remains a tool of justice rather than a substitute for constitutional judgment.
About the Author
Vridhi Parakh is a second-year student currently pursuing B.A LL.B (Hons) at O.P Jindal Global University. Her interest lies in exploring the various facets of criminal law, criminology and the forensic sciences.
Image Source:
AI and Privacy Law in India: Navigating the Digital Transformation Through Legal Innovation