The Black Box Counsel: A Study On The Admissibility Of Opaque Algorithmic Outputs Under The Indian Law Of Evidence And The Delimitation Of Delictual Liability For Computational Legal Malpractice

Full Text Link

Aditya Ballolli, School of Law, Christ University1

ABSTRACT

The integration of sophisticated, proprietary Artificial Intelligence (AI) into legal practice is poised to redefine the Indian legal landscape, offering unprecedented efficiencies in legal research, predictive analytics, and litigation strategy. However, the deployment of these “black-box” AI tools, whose decision-making logic is opaque, presents a profound challenge to the foundational human-centric principles of India's legal system. This article investigates the critical regulatory void surrounding the use of such AI by legal practitioners, a domain currently unaddressed by existing Indian law. It argues that the Advocates Act, 1961, the Indian Evidence Act, 1872, and the Bar Council of India Rules are fundamentally ill-equipped to handle the dual crises of admissibility of AI-generated legal work product and the allocation of liability for algorithmic error. This article critically analyses how the opaque nature of proprietary AI conflicts with core evidentiary principles, such as the right to cross-examine expert testimony under Section 45 of the Evidence Act (Section 39 in the new Bharatiya Sakshya Adhiniyam (BSA). Simultaneously, it deconstructs the emerging liability conundrum, questioning how tort law principles of professional negligence can be recalibrated when legal malpractice stems from a lawyer's reliance on flawed AI output. By comparing India's regulatory silence with nascent regulatory discourses in jurisdictions like the United States and the European Union, this article highlights the urgent need for a tailored Indian response. Looking forward, it proposes a pioneering framework for “AI Legal Practice Guidelines” to be adopted by the Bar Council of India, establishing standards for technological competence, duty of verification, and a test for apportioning liability. Through this exploration, the article aims to provide an essential scholarly foundation for policymakers, legal professionals, and scholars to navigate the complexities of integrating automation without compromising ethical accountability, fair trial standards, and the integrity of the legal profession in India.

343. THE BLACK BOX COUNSEL- A STUDY ON THE ADMISSIBILITY OF OPAQUE ALGORITHMIC OUTPUTS UND

Download THE BLACK BOX COUNSEL- A STUDY ON THE ADMISSIBILITY OF OPAQUE ALGORITHMIC OUTPUTS UND • 349KB