TRAI AI Guidelines for Telecom: Draft Rules for Jio, Airtel, Vi, BSNL and What Changes for Users
India’s telecom regulator has proposed the first dedicated AI rulebook for telecom companies. TRAI released draft TRAI AI guidelines for telecom on 30 October 2026 — “if a telecom company uses AI in its network or services, it must answer for what that AI does.”
Quick facts
- What: Draft “Guidelines on Responsible Use of Artificial Intelligence in Telecommunication Networks and Provision of Telecommunication Services”
- Released: 30 October 2026, for public consultation
- Comments deadline: 27 November 2026
- Who it covers: Jio, Airtel, Vodafone Idea (Vi), BSNL and other licensed telecom operators and network companies that use, or plan to use, AI
- Status: Draft only — not a final rule yet, and it does not change call rates, data tariffs, SIM rules or recharge plans
Why TRAI Is Making AI Rules for Telecom
Telecom companies already use AI in network management, customer service chatbots, fraud detection and service quality. Without checks, that AI can fail in ways users cannot see or challenge. The draft applies the government’s India AI Governance Guidelines (February 2026) to telecom, built on seven principles: trust, people first, innovation, fairness, accountability, understandable design, and safety and sustainability.
Who Is Covered by the Draft
The draft covers any entity holding a telecom licence or authorisation to provide telecom services, or to establish, operate, maintain or expand networks, if it uses or intends to use AI.
What TRAI Has Proposed: The Key Rules
1. A board-approved AI policy
Every operator using AI must have a board-approved AI policy, reviewed at least once every financial year.
2. Risk classification for every AI system
Each AI system must be classified as high, medium or low risk, based on its importance to the network and services, its financial and operational impact, and its effect on consumers.
3. Approval before use, and human oversight
AI systems need approval from a designated officer, with validation, conformity and impact checks before deployment or any change. Human oversight must stay in place, and data must be handled securely and lawfully.
4. An AI Governance and Safety Committee
Operators must set up an AI Governance and Safety Committee, led by a senior executive, to oversee how AI is used.
5. Incident reporting to TRAI
All AI incidents must be recorded and investigated, and critical- and high-severity incidents must be reported to TRAI. Operators must keep central databases — one for AI incidents, and one for reports of AI-related harm and their resolutions.
6. Accountability stays with the operator
The operator remains accountable even when a third-party vendor’s system is involved. The draft sets no separate penalty schedule — violations fall under existing laws and licence conditions.
What This Means for Mobile Users
Nothing changes immediately. This is a consultation draft, not a new charge or a SIM rule. If it is finalised, users should see fewer unexplained AI failures, with testing, records and reporting of serious failures, and a clearer complaint route when an AI-driven decision affects them — for example a blocked service, a wrong charge or a failed verification.
This is a sensible first step, but a draft is only as strong as its final version. Whether TRAI keeps the reporting duties strict after industry feedback will decide whether users actually feel the difference.
What Happens Next
Written comments are invited by 27 November 2026. Comments can be sent to TRAI Advisor (Networks, Spectrum and Licensing) Akhilesh Kumar Trivedi. After reviewing the feedback, TRAI will decide its next steps before finalising the rules. We will update this post when the final rules are announced.
Related reading: TRAI Mandated Packs Recharge Rule — our earlier post on TRAI’s recharge-pack rule.
Do you think telecom companies should be answerable when their AI systems fail? Tell us in the comments.
