New Delhi: The regulator has put the question on paper before the answer arrives on the network.
The Telecom Regulatory Authority of India on Friday released draft guidelines for the responsible use of artificial intelligence in telecommunication networks and in the delivery of telecom services, and invited public comments until November 27, 2026.
The draft does not invent a new framework. It takes the India AI Governance Guidelines, which the government released in February 2026, and applies them to a sector that TRAI says carries special weight.
"The telecommunication sector constitutes a significant part of India's digital infrastructure," the Ministry of Communications said in a statement. The ministry noted that the sector supports connectivity for the public, delivery of public and essential e-services, trade and commerce, and "secure and resilient functioning of the country's critical information infrastructure".
The seven principles the draft follows are: Trust is the Foundation, People First, Innovation over Restraint, Fairness and Equity, Accountability, Understandable by Design, and Safety, Resilience and Sustainability.
In plain terms, the draft covers much of what already happens quietly on Indian networks. AI is used to predict equipment failure, reroute traffic, flag unusual activity and handle customer support requests. When a model misjudges a fault during a cyclone, or wrongly flags a number as fraudulent, the decision lands on a subscriber who never spoke to a person.
As released, the draft proposes no new penalties, no mandatory AI audits and no immediate compliance deadlines. TRAI has said it will frame the guidelines after consultation. Comments are to be sent to Akhilesh Kumar Trivedi, Advisor for Networks, Spectrum and Licensing, at [email protected].
The four big operators — Reliance Jio, Bharti Airtel, Vodafone Idea and state-run BSNL — are the companies most directly in scope, though the guidelines cover, more broadly, every service provider that uses AI in network and service delivery.
None of the four had issued a public response to the draft at the time of filing.
The consultation deadline gives the industry roughly four weeks to argue about how much of the model has to be explainable, who signs off on an automated decision, and what a subscriber can do when an algorithm, not a person, has cut them off.
That last question is what the paper is really about.