TRAI drafts AI rules for telecom networks

What TRAI has put out for comment

27 November 2026 is the last date for written comments on a draft that could shape how artificial intelligence is governed inside telecom networks. The Telecom Regulatory Authority of India (TRAI) released the draft Guidelines on Responsible Use of AI in Telecommunication Networks and Provision of Telecommunication Services on 30 October 2026. TRAI's "What's New" page lists the draft release dated 30/10/2026, and a report on the release was published at 05:32 PM IST the same day. The document is a draft proposal, not final rules in force. That distinction matters for every duty described below. Nothing in the draft coverage says these requirements already apply, and readers should not treat the draft as settled law. TRAI has anchored the draft in the Centre's India AI Governance Guidelines, released in February 2026. The draft draws on seven principles from that framework: Trust is the Foundation; People First; Innovation over Restraint; Fairness and Equity; Accountability; Understandable by Design; and Safety, Resilience and Sustainability.

What the draft would require

Under the proposal, covered entities would need a board-approved AI policy. They would also need a committee led by a senior AI Governance and Safety Officer. That pairing is the centre of the governance model in the draft: a written policy approved at board level, and a named senior officer leading the committee that oversees it. The draft also proposes risk classification of AI systems as high, medium or low, and an inventory of AI systems. Human oversight is part of the proposal as well. In addition, critical- and high-severity AI incidents would have to be reported to TRAI. No reporting deadline, prescribed reporting form, or penalty for late reporting is stated in the coverage of the draft, so none is given here. The practical effect, if the draft were adopted in this form, would be a move from general principles to documented duties. An entity would need to be able to show what AI systems it uses, how it has classified their risk, who oversees them, and what it did when a serious incident occurred. No individual company is named in the coverage of the draft.

What changed

Before this draft, the February 2026 India AI Governance Guidelines set the broader framework, without telecom-specific duties attached to it. TRAI's 30 October 2026 draft is the step that would translate that framework into proposed duties for telecom networks and telecom services. That is the change to watch. The question is no longer only whether AI should be used responsibly in telecom. The draft asks how that responsibility should be organised, recorded and reported, and who inside an entity should answer for it.

Who is affected and what to do

Telecom entities covered by the draft would face the most direct work. On the proposal as described, they would need to prepare or update a board-approved AI policy, set up the senior-officer-led committee, build and maintain an inventory of AI systems, apply the high, medium and low risk classification, put human oversight in place, and prepare a route for reporting critical- and high-severity incidents to TRAI. Industry bodies, companies, researchers and other interested readers who want to shape the final version can submit written comments by 27 November 2026. TRAI's release names Shri Akhilesh Kumar Trivedi, Advisor (Networks, Spectrum and Licensing), as the comments contact. No comments email address or portal is given in the coverage available, so none is listed here. Readers who plan to comment should check TRAI's release itself for the submission route before the deadline. No action for subscribers is described in the draft coverage. The duties in the draft sit with the covered entities, not with individual users. A subscriber would therefore not need to file anything or change a service because of this draft alone. Any change a subscriber might notice later would depend on what TRAI finally adopts, and that outcome is not yet known.

Key terms in plain language

AI policy: a written, board-approved statement of how an entity will govern its use of AI. AI Governance and Safety Officer: the senior officer who would lead the committee overseeing AI governance and safety under the draft. Risk classification: sorting AI systems into high, medium or low risk so that oversight matches the level of risk. Inventory of AI systems: a recorded list of the AI systems an entity uses, so they can be tracked and reviewed. Human oversight: a person, rather than an automated system alone, remaining responsible for checking and controlling AI use. Incident reporting: telling TRAI about critical- and high-severity AI incidents, as proposed in the draft.

Timeline

February 2026: the Centre releases the India AI Governance Guidelines. 30 October 2026: TRAI releases the draft Guidelines on Responsible Use of AI in Telecommunication Networks and Provision of Telecommunication Services. 27 November 2026: last date for stakeholders to submit written comments. After that date, the next step is TRAI's consideration of the comments received. No date has been announced for any final guidelines, so none is given here.

What happens next

The comment window runs until 27 November 2026. That is the immediate date to watch. Once comments close, the next question will be whether TRAI revises the draft, and in what form, before any final guidelines are issued. Until TRAI issues a final version, the draft remains a proposal, and the duties described above remain proposed duties, not rules in force. 📷 Illustration: AI-generated