New Delhi: The experimental use of Artificial Intelligence (AI) has increased in the legal sector, including within courts, according to a new report. It noted that although AI use is compelling in an overburdened system like India’s judiciary, caution is necessary because some judicial officers, law clerks, and secretarial staff engage in possible “shadow use” of AI, without any institutional safeguards in place.
The DAKSH and Digital Futures Lab report, ‘AI for Justice: Ethical, Fair and Robust Adoption in India’s Courts’, was released Tuesday at the United Nations House in New Delhi.
The report cautioned against the unrestricted use of AI. “In the absence of expertise, training or guidelines for use, many judges and court staff are experimenting with primarily free tools, often unaware of the associated risks, such as the potential for AI tools to hallucinate precedents,” said the 82-page report launched in collaboration with the United Nations Development Programme (UNDP).
The report also highlighted incidents where lawyers erroneously cited fabricated or false cases while arguing before Indian courts, which are often attributed to AI.
Founded in 2015, DAKSH is a civil society organisation that has been working on judicial reforms for over a decade now. The UNDP commissioned the report under the project ‘Strengthening Rule of Law, Access to Justice pathways to accelerate India’s Sustainable Development’.
Delhi High Court Chief Justice D.K. Upadhyay’s keynote address at the event emphasised that AI is not just being adopted across Indian courts but is here to stay. “A robust governance framework for AI in the judiciary is necessary for courts to adopt such technology in a manner that is balanced, fair, and tailored towards meeting tasks that require AI assistance,” Justice Upadhyay told the audience composed of lawyers, judges, and UN representatives.
The report’s launch included a panel discussion featuring former Chief Justice of the Himachal Pradesh High Court Rajiv Shakdher; Executive Director of the Centre for Communication at National Law University Delhi, Jhalak Kakkar; and UNDP’s Regional (India) Representative Angela Lusigi.
Although AI can support courts in working more efficiently and effectively, Lusigi said, “Such use should be approached with care, keeping in mind the concerns of transparency and respect for rights.”
“We hope that this report and assessment frameworks help courts make informed choices about what use cases are amenable to AI use, how to select AI tools and what safeguards to put in place while using AI,” she added.
One of the report’s authors, DAKSH Research Manager Leah Verghese, told ThePrint, “Indian courts can no longer ignore AI, but its use in judicial contexts need to be approached with caution. Given the nature of AI, it has the potential to improve the delivery of justice and also deepen existing barriers to accessing it.”
Also Read: Deepfake on duty: when I asked AI to read Op Sindoor citations
How AI is being used in Indian courts
One of the most significant findings noted in the report was that the use of “shadow AI”—as in other workplaces—has been rising in courts.
One of the major ways in which AI is being used in Indian courtrooms and, generally, by lawyers and interns in India’s legal sector is drafting. This could entail drafting case briefs, contracts, documents, orders, judgments and decrees, among others. Prior to this, drafting was done manually by judicial officers with support from law clerks and secretarial staff.
AI’s promise of convenience, time-saving, and enhanced support is particularly compelling in a system as overburdened as the Indian judiciary, the report said. It noted that troubles faced by courts, such as mounting pendency figures, could be lessened with the use of AI tools.
As of 15 December 2025, nearly 5.5 crore cases were pending in India’s district and high courts, which strengthened the case for AI use.
A 2024 global UNESCO survey revealed that 44 percent of (India’s or world’s???) judicial operators used AI tools for work-related tasks, according to the report. Against this backdrop, the report highlighted the need for caution.
Increasing use of AI in India’s legal sectorAs the world’s second-largest legal-tech market, AI is increasingly being seen in India as a “force for good”, with AI-enabled tools being rolled out in courts, the report said.
For instance, in October 2025, the Kerala High Court directed all courts in the state to adopt Adalat AI, an automated transcription tool for recording witness depositions.
The report, however, cautioned that the use and deployment of AI among Indian courts remained ad hoc and poorly documented. AI use is most visible in tasks such as court proceedings’ translation and transcription, according to the report.
Underlining that digitisation efforts such as the eCourts venture adopted by the Indian courts—aimed at bringing about technological communication and management-related changes to the Indian judiciary—had been “substantial”, the report added that such initiatives had been adopted mainly by the Department of Justice, the Supreme Court eCommittee, HCs, and the National Informatics Centre (NIC).
Risks of using AI in legal settings
The report explained how AI systems operate through data, models and outputs—each of which comes with concerns relating to accuracy, fairness, and privacy in sensitive cases. As custodians of highly sensitive information, courts must exercise caution in sharing data with external actors who do not operate under the same ethical mandates or levels of public responsibility, it said.
“Risks such as bias, hallucinations, lack of transparency, and limitations in underlying datasets pose specific threats to judicial legitimacy,” the report said.
It added that many courts lacked proper technical capabilities and impact assessment processes, which in turn made it difficult to evaluate whether AI integration was actually improving judicial efficiency.
The report said that since the use of AI was still largely driven by individuals—not organisations—the process was vulnerable to disruption due to judicial transfers or retirements of the individuals who spearheaded its use. Concerns surrounding overstretched administrative capacity and security, accuracy, and ethics also remained, it added.
“Taken together, these gaps risk eroding public trust in courts and raise concerns about the legitimacy of the judiciary as AI use expands,” the report pointed out, while adding that AI use also came with the potential violation of rights.
As India’s legal-tech market expands and courts increasingly experiment with AI in judicial processes, now is an opportune time to develop decision-making models for designing and adopting AI systems in courts, according to the report. It pointed out that in this day and age, day-to-day judicial work are still heavily reliant on physical files, paper registers, and manual processes.
(Edited by Madhurita Goswami)