AI in the Judiciary: Supreme Court’s Draft Framework for Responsible AI Governance

AI in the Judiciary Latest News

• The Supreme Court, in June 2026, released the draft Regulations for Use of Artificial Intelligence (AI) in Courts, 2026, aimed at creating a governance framework for AI adoption in the judiciary. 
• It prescribes general principles for AI’s use and establishes an institutional framework to oversee it.

Applicability: Is It Binding?

• The Draft Regulations are not automatically or uniformly binding. 
• They will come into force for the Supreme Court on a date notified by the Chief Justice of India, and separately for each High Court (and courts/tribunals under its jurisdiction) on dates notified by that High Court’s Chief Justice. 
• Different provisions can also be implemented on different dates, allowing phased AI adoption suited to each court’s needs.

‹ ›

What Role Has Been Carved Out for AI?

• Courts are required to “actively seek opportunities” to deploy AI systems that demonstrably improve access to justice, reduce delays, or enhance administrative efficiency. 
• AI use is explicitly permitted for administrative and assistive functions such as case management, transcription, translation, legal research, document summarisation, accessibility, and court administration. 
• All such use requires written approval from the SC’s Apex Body or the concerned High Court/tribunal’s AI Committee, along with supervision by nominated court officers.

Can AI Decide a Case?

• No. The Draft Regulations categorically state that no judicial outcome can be reached through algorithmic decision-making alone or solely on AI-generated information. 
• Human judicial authority remains determinative in all adjudicative decisions; AI’s role in decision-making, if any, is purely advisory and subject to independent human evaluation.

What Is Absolutely Barred?

• Certain uses are prohibited in “absolute and non-derogable” terms, meaning no authority can permit them under any circumstances:
  • Risk scoring to assess flight risk;
  • Predicting recidivism (the process of using historical, behavioral, and demographic data to estimate the likelihood that a previously convicted individual will commit a new crime or return to prison); 
  • Evaluating bail eligibility;
  • Determining witness credibility;
  • Predicting, profiling, or inferring future conduct of parties, accused persons, witnesses, or legal representatives;
  • Submitting AI-generated output as independent evidence without full disclosure of its AI origin;
  • Using “black-box” (unexplainable) AI systems in matters affecting personal liberty.

Transparency for Litigants

• If a court uses AI that “materially assists” it in case management, document analysis, or judicial administration, it must inform the parties involved in a timely, accessible manner. 
• However, litigants will only be notified when AI has provided material assistance, not for every instance of AI use.

Institutional and Regulatory Architecture

• An “Apex Body” at the Supreme Court, comprising sitting SC and High Court judges, a MeitY official, and finance/cybersecurity experts, will set minimum mandatory standards and issue implementation guidelines.
• This body will operate through five specialised committees, while the SC and each High Court will form their own AI Committees, backed by an AI Secretariat.
• A separate research body, the Centre of Research and Excellence on Artificial Intelligence (CoRE-AI), will evaluate tools and track developments to support the Apex Body.

Role of Private Vendors

• Private companies can be involved only with written approval from the relevant court authority and must comply with mandatory contract terms covering:
  • Ownership of and access rights to court data and AI outputs;
  • A bar on using sensitive judicial data;
  • A prohibition on retaining or fine-tuning models using court data without AI Committee approval;
  • No exclusive IP claims over tools built substantially on judicial data or public resources.

Safety Measures: A Lifecycle Approach

• The Regulations mandate safety checks before, during, and after AI deployment:
  • A Technical and Ethical Impact Assessment covering system architecture, training data, bias, hallucination risks, and cybersecurity
  • Some systems may require “Controlled Environment Testing” in isolation before deployment
  • Post-deployment, systems face technical, legal, ethical, and cybersecurity audits conducted in-house (source code and training data cannot be shared with third parties)
  • Each court must maintain an AI Register of approved systems and audit outcomes
  • Each AI Secretariat must maintain an AI Incident Database, with mandatory 24-hour notification if a tool fails or is suspended
  • High Courts must have an emergency fall-back protocol to run processes manually if systems fail

Grievance Redressal for Litigants

• Where harm results from a prohibited use of AI, an affected party can file an application with the court where the AI system was used. 
• The court must provide a hearing and pass appropriate orders. 
• High Courts may design simplified procedures and complaint formats for accessibility, and this remedy exists alongside other legal remedies already available under ordinary law.

Conclusion

• The SC’s draft regulations strike a calibrated balance, embracing AI for efficiency and access to justice while firmly safeguarding judicial independence, personal liberty, and due process. 
• By barring algorithmic decision-making in sensitive matters and ensuring layered oversight, India moves toward responsible, human-centric AI governance in courts.

Source: IE | TLL