The evolution
Law has absorbed technology before — the typewriter, the fax, e-filing, video hearings. In July 2026 the Supreme Court of India said as much: dependence on technology has never troubled the dispensation of justice.
AI breaks the pattern. It is not a faster pen. It substitutes for the reasoning the pen was only recording. Retrieval came first, then prediction and automation. Since late 2022, generative models produce legal text indistinguishable in form from legal work. A hallucinated citation looks exactly like a real one. The technology can counterfeit the output of legal reasoning with none of the reasoning. Everything that follows is a response to that single fact.
Across the fields of law
Transactional practice uses AI for diligence and clause extraction — low-risk work verifiable within the four corners of a document. Criminal law raises the sharpest civil-liberties questions: predictive policing, facial recognition, bail risk scoring. IP practice uses it for prior art and drafting. Access to justice may gain the most, through translation and plain-language procedure.
The pattern is simple. AI performs best where the ground truth sits inside the document and worst where it sits outside — in a law report, in the state of the art, in the world. Summarisation is safe. Assertion is dangerous.
The Indian landmark
In Pooja Ramesh Singh v. Jammu and Kashmir Bank Ltd., 2026 INSC 668 (2 July 2026), the Supreme Court set aside NCLT and NCLAT orders because the tribunal had relied on six precedents that either did not exist or did not say what they were said to say. The fakes came from the tribunal's own research, not from counsel. A decision resting on fabricated material is no decision in the eyes of the law, even where only an iota of such material entered. Contamination is enough.
The Court declared zero tolerance for Bar and Bench alike, held that citing such material unverified is professional misconduct, and directed the Bar Council of India to prescribe norms. It did not retreat from technology. It asserted total control over adjudication, with a human in the loop at every stage.
The Draft Regulations, 2026
On 3 June 2026 the Supreme Court's AI Committee released the draft Regulations for Use of AI in Courts, 2026 — 57 regulations built on human primacy, explainability, auditability and proportionality. Regulation 19 permits case management, transcription, translation, research and anonymisation. Regulation 20 prohibits absolutely: outcomes by algorithm alone, risk scoring for bail or credibility, behavioural profiling, and surveillance of judges or litigants. Regulation 43 requires a party using AI to prepare a pleading to declare it.
The Supreme Court of Victoria's guidelines of 14 May 2026 run to two pages. India builds an institution; Victoria states a norm. Both land on the same line.
Patent law
Patent law is where AI becomes existential, because its central concept is the act of having an idea. Inventorship is now settled. Thaler failed in the United States, the United Kingdom, Europe and Australia, and India refused the DABUS application on 15 April 2026 — a machine is not a “true and first inventor”. On 28 November 2025 the USPTO rescinded its 2024 guidance and now treats AI as laboratory equipment.
The quiet question is obviousness. As practitioners routinely deploy AI, ordinary skill rises to include the machine's knowledge. A person skilled in the art who has read everything finds everything obvious. Either everything gets patented, or nothing does.
The assistive/substitutive line will hold, because it is the only line that preserves accountability. Verification is a non-delegable professional duty.
Within a few years, an AI-assistance declaration will likely accompany filings as routinely as a vakalatnama.
TECHNOLOGY IS WELCOME. RESPONSIBILITY IS NOT TRANSFERABLE.
References
Pooja Ramesh Singh v. Jammu and Kashmir Bank Ltd., 2026 INSC 668 (SC, 2 July 2026).
Supreme Court of India AI Committee, [Draft] Regulations for Use of AI in Courts, 2026 (3 June 2026).
Supreme Court of Victoria, Use of AI by Judicial Officers (14 May 2026).
Henry H. Perritt Jr., AI Co-Inventor: Robot Enabled Patent Prosecution, 25.2 UIC Rev. Intell. Prop. L. 1 (2026); U.S. Patent No. 12,409,564.
Tim W. Dornis, Artificial Intelligence and Innovation: The End of Patent Law As We Know It, 23 Yale J.L. & Tech. 97 (2020).
USPTO, Revised Inventorship Guidance for AI-Assisted Inventions (28 Nov. 2025), rescinding 89 Fed. Reg. 10043 (2024).
Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022); Thaler v. Comptroller-General [2023] UKSC 49; EPO J 8/20.
Indian Patent Office orders dated 15 April 2026, Application No. 202017019068.
Kerala High Court, Policy on Use of AI Tools in District Judiciary (19 July 2025).
Regulation (EU) 2024/1689 (AI Act); Regulation (EU) 2026/1744 (Digital Omnibus on AI).
Disclaimer
For information and academic discussion only; not legal advice, and no advocate–client relationship arises. As at the date of writing, the Regulations for Use of AI in Courts, 2026 remain a draft under consultation and not notified law; the BCI committee had not published norms; and the USPTO's revised guidance is examination policy, not binding law.
","deck":"","published_at":"2026-09-03T04:27:36.403Z","title":"MASTER AND ALGORITHM","url":"https://kandulamahesh.com/insights/master-and-algorithm"}