5 Key Takeaways
- SCAORA rejects mandatory AI disclosure, arguing it is unnecessary and potentially prejudicial.
- Existing professional obligations under the Advocates Act already ensure accuracy, making disclosure redundant.
- Mandatory disclosure could bias judges and litigants against AI-assisted documents, harming fairness.
- SCAORA proposes an amended certificate requiring personal verification of citations instead of AI use disclosure.
- The association calls for caution, audits of current AI tools, inclusive governance, and clear definitions to prevent risks like automation bias and hallucinations.
India's Top Lawyers' Body Rejects Mandatory AI Disclosure in Supreme Court Draft Rules
SCAORA pushes back against proposed regulations, calling the disclosure mandate "unjustified" and warning of prejudice against litigants who use AI-assisted drafting tools.
A quiet but significant battle over the future of artificial intelligence in India's courtrooms is unfolding, and at its heart lies a simple question: should lawyers be forced to tell a judge every time they use an AI tool to help draft a legal document?
The Supreme Court Advocates-on-Record Association (SCAORA), the powerful body representing the lawyers authorized to practice before the country's highest court, has delivered an emphatic answer — no. In a detailed submission to the Supreme Court's AI Committee, SCAORA has demanded the removal of a proposed rule that would make such disclosure mandatory, arguing it is both unnecessary and potentially harmful to the fairness of legal proceedings. The demand is one of many that pushes back against what the association sees as a rushed and less-than-inclusive approach to regulating artificial intelligence across India's judicial system.
On the surface, the draft reads like a cautious embrace of technology. It gives a green light to using AI for legal research, drafting assistance, language translation, real-time transcription, case management, scheduling, and even accessibility services for disabled court users. However, it draws a bright red line: AI cannot decide a case, pass a sentence, determine bail, judge the credibility of a witness, or replace human judicial decision-making in any core adjudicatory function.
Buried within that framework, though, are two clauses — Regulations 43(3) and 43(4) — that would compel any party to a legal proceeding to disclose whenever AI tools have been used in the preparation of pleadings, documents, or evidence. It is this disclosure mandate that has drawn the sharpest fire from SCAORA.
The association's central contention is that under the Advocates Act, 1961, and the Supreme Court Rules, 2013, an advocate is already personally responsible for the accuracy and integrity of every single document they file before a court. Lawyers are officers of the court, bound by a strict statutory code of professional conduct. If an AI tool produces a hallucinated case citation or a flawed legal argument, it is the lawyer — not the machine — who faces professional embarrassment, judicial rebuke, or disciplinary action. To pile a separate AI disclosure requirement on top of that existing liability, SCAORA argues, is redundant.
The association went further, warning that the practical consequences of routine disclosure could be deeply unfair. Requiring a lawyer to state on the record that AI was used could, in the words of SCAORA, "prejudice litigants" and even "influence the presiding judge." A judge might unconsciously give less weight to a pleading that bears an AI-usage label, or a litigant on the other side might try to spin the disclosure as a sign of a weaker case, undermining the principle that arguments should be judged on their merit, not on the tools that shaped them. In a legal culture that is still learning to trust the role of machines, the mere mention of AI involvement could tip the scales in unpredictable ways.
In place of the mandatory disclosure, SCAORA has offered a surgical alternative. Rather than forcing lawyers to declare every use of a drafting assistant, they propose amending the existing certificate that every advocate must attach to a Special Leave Petition — the gateway document for many cases at the Supreme Court. The amended certificate would require the advocate to swear that all legal citations, judicial precedents, and statutory authorities relied upon in the document have been personally verified. This keeps the focus squarely where it belongs, SCAORA believes: on the accuracy of the final product, not on the process that helped create it.
But the pushback against the draft regulations goes far beyond the disclosure clause. SCAORA's submission is a sweeping call for caution, inclusivity, and a fundamental rethinking of the speed at which India's courts are running toward AI adoption. The draft regulations, as written, emphasize "innovation over restraint." SCAORA wants that phrase turned on its head. The association is advocating for a precautionary approach, especially when AI is deployed in functions that directly affect the rights of litigants. Until adequate safeguards are in place, any move that touches core legal rights should be paused, it says.
The association flagged a trio of dangers that anyone who has used generative AI will recognize:
Another major fault line running through SCAORA's response is the question of who gets to design and oversee this brave new AI-powered judicial world. A close look at the governance framework proposed in the draft regulations reveals, in SCAORA's view, a glaring omission: the voices of practicing advocates have been largely excluded. The proposed governance structure includes a sprawling network of bodies — an Apex Body, a Judicial Committee, a Case and Data Management Committee, and multiple AI Committees — but the lawyers who appear day after day in India's courtrooms, who will be among the primary users of these AI tools, have been left without a meaningful seat at the table.
SCAORA called this exclusion unacceptable and proposed a direct fix: make the President of SCAORA, or a nominee, an ex-officio member of the Apex Body, and ensure that practicing advocates are represented on the subordinate committees that will shape how data is managed and which AI tools are approved.
The governance critique extends to structure and transparency. SCAORA described the proposed architecture as excessively bureaucratic, bogged down by overlapping committees and convoluted approval mechanisms that could stifle the very innovation the regulations seek to channel. Adding to the discomfort are data protection concerns. The draft regulations grapple with the sensitive matter of "Court Data" — a term that could encompass everything from a litigant's personal details to the confidential legal strategies of government bodies. SCAORA's submission raises a red flag: without robust data protection safeguards, the rush to digitize and automate could turn court records into a privacy liability.
Before any new framework is finalized, SCAORA insists, the Supreme Court must first look inward at the AI tools it has already deployed. The court has not been sitting idle; a suite of AI-powered systems is already operational under its own roof:
AI Systems Currently Operational at the Supreme Court
- SUPACE — Supreme Court Portal for Assistance in Court's Efficiency; intelligent case information retrieval for judges
- SUVAS — Supreme Court Vidhik Anuvaad Software; machine translation tool for judicial documents
- SuSahayak — Voice-to-text transcription service
- AI-Assisted Transcription — Converts live court proceedings into text
- e-SCR — Digital platform for accessing Supreme Court records
- AI-Enabled E-Filing — Streamlines case submissions
SCAORA's demand is unambiguous: each of these systems must undergo a rigorous technical, legal, ethical, and cybersecurity audit before the court takes another step forward with a broad AI regulatory framework. To do otherwise would be to build a house on an unexamined foundation.
Compounding the challenge is what SCAORA sees as a fog of undefined terms that could leave the regulations dangerously open to interpretation. The submission calls for crystal-clear definitions of foundational concepts. What exactly falls under "Court Data"? Who counts as an "AI Service Provider" — the software company, the cloud provider, the consultant fine-tuning the model? How does one classify a tool as a "High-Risk AI Tool"? And what precisely is meant by "judicial resource allocation"? Without precise definitions, SCAORA warns, the regulations risk becoming either unenforceably vague or arbitrarily applied.
Throughout its submission, SCAORA returns to a single refrain: AI must remain assistive. The draft regulations, for all their permissive clauses, already state this prohibition clearly — AI shall not decide cases. SCAORA's push is to ensure that this prohibition is not diluted in spirit by a thousand small delegations to machines, and that the systems built around the courts respect the profession that serves them.
The Supreme Court's AI Committee has thrown open its doors for comments and suggestions before finalizing the framework. The SCAORA intervention, with its blend of professional ethics, technical prudence, and institutional demand for representation, is likely to carry significant weight. What happens next will shape not just how lawyers work, but how justice itself is administered in an age where the line between human judgment and machine output grows thinner by the day. For now, the ball is in the Supreme Court's court, and the legal fraternity is watching closely to see whether the final regulations will compel disclosure, mandate audits, open governance doors to practicing advocates, and ultimately, heed the call to move fast — but not too fast — with AI.
No comments:
Post a Comment