India’s Trade Marks Registry proposes mandatory AI disclosure, bars AI-generated affidavits and altered evidence

Filing before India’s Trade Marks Registry will now require everyone to declare whether they have used artificial intelligence (AI). The Office of the Controller General of Patents, Designs and Trade Marks (CGPDTM) has published draft guidelines in this regard, which also bar the use of AI to write affidavits or alter evidence, including labels, invoices and advertisements.

The draft applies to every applicant, agent and advocate appearing before the Registry, including in pending matters. It makes the person filing a submission responsible for every case law citation. The guidelines are not currently in force and will take effect once the office notifies them.

Separately, in a public notice dated October 16, the CGPDTM’s office invited suggestions on guidelines to address the “use/misuse of Artificial Intelligence (AI) and the menace of squatting”. Squatting refers to the practice of registering a mark used by another party, usually to block the brand owner or sell the mark back to them. The notice follows an October 13 Delhi High Court order in Stryten Energy LLC v. Nidhi Agarwal & Ors., which stated that “there has to be comprehensive guidelines” on both issues.

What does the draft require?

• A declaration with each filing: Each submission must include information on “whether or not the submission has been prepared with the assistance of an AI”. The draft treats non-disclosure as “a misrepresentation to the Registry”.
• Human checks: Any content suggested by AI for drafts, particularly citations and facts, must be verified against sources such as court websites, the Official Gazette, certified copies or legal databases. An AI-generated citation does not count as verified “unless independently cross-checked”.
• No AI in sworn statements: The draft bars the use of AI to “generate, author, embellish, strengthen, dilute, fabricate, or otherwise alter” the substance of an affidavit, declaration or similar statement.
• No AI-altered evidence: AI also cannot be used to “generate, fabricate, materially manipulate, or digitally alter” supporting material, including any “image, label, invoice, sales data, market survey, website extract, advertisement” or “translation”. If the Registrar suspects that AI has been used, they may demand “the original or native file” and its metadata.
• Penalties: After issuing a show-cause notice and holding a hearing, the Registrar may strike off a false citation or the entire submission. Agents may also be referred to a disciplinary committee, while advocates may be referred to their respective Bar Councils. The citation rule applies “regardless of whether such incorrect citation originated from an AI”.

What does the draft leave out?

The public notice does not mention the draft guidelines. The draft itself states that it was prepared pursuant to a prior order in the same case, dated April 16, 2026. It also addresses AI but does not deal with squatting.

The draft mentions “Examiners/Registrar” in the context of the misuse of AI-generated content and asks officers to keep AI “strictly ancillary”. However, its penalties apply only to parties and their representatives.

There have also been instances of adjudicators relying on fabricated judgments. On July 2, 2026, the Supreme Court of India set aside an order of the National Company Law Tribunal and its appellate tribunal for relying on non-existent, AI-generated precedents, as reported by LiveLaw.

The draft also states that it will be read in harmony with the “Supreme Court Artificial Intelligence Regulations, 2026”. However, MediaNama could not establish whether these regulations had been notified. The Supreme Court’s AI Committee had published the draft Regulations for Use of AI in Courts on June 3, with no subsequent updates that MediaNama could find.

Questions MediaNama has sent to the CGPDTM

MediaNama has sent the following questions to the CGPDTM’s office and will update this story if it responds:

• Will the office publish a draft on squatting before reporting to the court?
• The draft defines AI to include “Natural Language Processing tools”. Would using a grammar checker or machine translation require a declaration, and would an AI-assisted translation breach the bar in clause 14.2 on AI altering a “translation”?
• Can a label or advertisement that a company actually uses, but designed with an AI tool, be filed as evidence that a mark is in use?
• Clause 8.3 asks for a declaration with every submission and clause 8.1 only where AI was used, while the form in Annexure I has no field for stating whether AI was used. Which rule applies?
• Does the Registry use AI tools in search, examination or hearings, and what follows if an officer relies on an unverified AI output?
• How many submissions has the office found citing non-existent case law, and over what period?

Also read:

• Lowdown: How the Supreme Court’s draft AI rules would govern Indian courts
• Gujarat HC releases AI policy barring use in court decisions, reasoning, and orders