Usage Of AI Can Save Time In Arbitration Proceedings, Cannot Be Ignored Over Confidentiality Concerns: Justice Vikram Nath

Speaking at the 'Arbitration 2.0: Integration of Artificial Intelligence and Other technologies to enhance the efficiency of arbitration' session organised by Delhi Arbitration Weekend 2025 (“DAW”), Supreme Court judge Justice Vikram Nath observed that law, despite not being a scientific field, does not discourage the legal fraternity from adopting artificial intelligence (“AI”).

The panel included Justice N. Anand Venkatesh, Judge, Madras High Court, Mr Tushar Mehtra, Solicitor General of India, Mr Mark Dempsey, SC, and Mr Tim Lord, KC, Brick Court Chamber.

Justice Nath observed that arbitration was adopted as an alternative to routine court proceedings as it is speedy and effective. However, arbitration cannot be stagnant and has to constantly evolve. Arbitration is being adopted by parties as a speedy dispute resolution, and AI has the potential to redefine what time even means in dispute Resolution.

Citing an example from the Singapore International Arbitration Centre (“SIAC”), a new gateway platform was launched to digitise and simplify arbitration case management. However, AI cannot be used as a substitute for arbitration to adjudicate disputes.

Justice Nath flagged that AI is not a replacement for human judgment but a powerful research assistant. AI is a translator that breaks down language barriers in cross-border disputes, an analyst who can help us see patterns we might miss.

Justice N. Anand Venkatesh spoke about how AI has started reading who we are and deciding what we want. He highlighted that AI can be used to identify the gaps in oral evidence and test arguments that are being put forth. However, the real challenge will be in the future when AI transforms from being a tool in the arbitral proceedings to an arbitrator adjudicating the dispute.

Mr Tushar Mehta observed that it is too early to take a call on the use of AI for the adjudicatory process. Wherever AI is used in any proceedings, it has failed miserably. It has been a victim of AI hallucination, i.e., creating fake judgments, losing track of the documents, etc. However, the real danger of AI is algorithmic bias.

Recently, in the US, AI was used to adjudicate disputes concerning traffic violations. The particular AI had a bias against coloured woman. Till the time we are sure of the efficacy of AI, it would be hazardous to rely on AI. The AI hallucination is one problem, but the AI Bias is the most problematic part. AI is an assimilation or collection of whatever material is available on the internet. Such material can be created, and therefore, a bias can be created.

Mr Mark Dempsey SC highlighted that AI can be used as a summarisation tool. However, the summary of proceedings or arguments has to be followed by the responsibility of the parties or the arbitrator using AI.

Mr Tim Lord KC observed that there is delegation in the human framework as well. A King's Counsel relies upon his junior to churn out the first draft. Therefore, the Arbitrator panel can also rely upon AI to delegate work. However, AI cannot be a substitute for the decision. It is the responsibility of the Arbitral Panel to have an oversight on the usage of the AI.