Last Updated:
SC issues draft AI rules for courts, allowing AI for research and administration but banning it from decisions, mandates disclosure, audits, data safeguards and human primacy
Supreme Court of India (PTI/File)
The Supreme Court has unveiled a comprehensive draft framework to regulate the use of Artificial Intelligence (AI) in India’s judicial system, making it clear that AI can assist courts but can never replace judges.
According to a report by Bar And Bench, the proposed Regulations for Use of Artificial Intelligence in Courts, 2026 allow AI tools to be used for legal research, drafting assistance, translation, transcription and court administration, while explicitly prohibiting them from making judicial decisions, determining bail, assessing witness credibility or passing sentences.
The draft regulations, prepared by the Supreme Court’s AI Committee, have been released for public consultation until June 20 and would apply to the Supreme Court, High Courts, subordinate courts, tribunals and statutory commissions across the country.
One of the key features of the framework is a mandatory disclosure requirement for lawyers and litigants. If AI is used to prepare pleadings, submissions, legal documents or evidence, the party must inform the court at the time of filing.
Courts will also be empowered to seek details about the AI platform used, the extent of AI assistance received and the verification process adopted to ensure accuracy.
The draft makes it clear that AI cannot be used as an excuse for mistakes. If AI-generated content is found to be false, fabricated or misleading, the responsibility will lie entirely with the person who submitted it.
AI Can Assist, Not Adjudicate
The Supreme Court has adopted what it calls the principle of “human primacy”, under which judges remain solely responsible for deciding questions of law, facts and justice.
Under the proposed rules, AI systems can assist with:
- Legal research and precedent searches
- Citation verification
- Summarising pleadings and judgments
- Translation of court documents
- Automated transcription of hearings
- Drafting assistance
- Case scheduling and cause list preparation
- Record management
- Administrative functions
- Accessibility services for persons with disabilities
- AI-powered litigant support systems
However, AI will be barred from:
- Deciding cases
- Determining judicial outcomes
- Passing sentences
- Deciding bail applications
- Assessing flight risk
- Evaluating witness credibility
- Predicting future criminal conduct
- Monitoring judges, lawyers or litigants
- Participating in judicial deliberations
The regulations also prohibit the use of opaque “black-box” algorithms in matters involving personal liberty or legal rights.
Supreme Court Recognises AI Hallucinations
The draft specifically addresses the growing concern around AI “hallucinations” — instances where AI systems generate convincing but inaccurate information.
The framework notes that AI tools may produce fabricated case laws, incorrect statutory provisions or misleading factual information. As a result, AI-generated outputs will remain advisory in nature and must be independently verified before being relied upon in judicial proceedings.
Judges and court officials will not be allowed to cite AI errors, hallucinations or system failures as a defence for incorrect or harmful decisions.
New AI Governance Structure
The Supreme Court has proposed a multi-layered oversight mechanism to regulate AI adoption in courts.
The framework calls for:
- A permanent apex AI authority at the Supreme Court level
- AI Committees in the Supreme Court and all High Courts
- Dedicated AI Secretariats
- A Centre of Research and Excellence on Artificial Intelligence (CoRE-AI)
These bodies would oversee approvals, audits, incident investigations, compliance monitoring and policy development.
Annual Audits And AI Registers
Every AI tool deployed in courts would be subject to technical, legal and ethical audits at least once every year.
Courts would also be required to maintain an AI Register documenting approved systems, audit findings, usage conditions and AI-related incidents.
An AI Incident Database is also proposed to record cases involving bias, system malfunctions, cybersecurity breaches and other technology-related failures.
Safeguards On Data And Private Vendors
The framework mandates compliance with the Digital Personal Data Protection Act, 2023 and places restrictions on how court data can be used.
Sensitive judicial information cannot be transferred to external AI systems without authorisation, while personal data cannot be used to train AI models without approval.
Private technology companies will also face strict conditions, including audit rights, cybersecurity obligations and restrictions on using judicial data for commercial purposes.
Courts Encouraged To Embrace AI
Despite the safeguards, the Supreme Court has adopted a generally pro-innovation approach.
The draft regulations establish a “presumption in favour of responsible AI adoption”, encouraging courts to use technology to improve efficiency, reduce case backlogs and enhance access to justice.
However, the framework draws a clear red line: AI may assist the judiciary, but the final authority to decide cases will always remain with human judges.
Delhi, India, India
Read More
Source link
[ad_3]
