India’s telecom regulator has proposed new governance rules for AI used by telecom operators and network providers, including risk classification, board oversight and incident reporting. If adopted, the rules would make the organisations using these systems responsible for tracking them and managing their risks, whether the AI is built in-house or bought from a supplier.
Watch Desk analysis
What happened
The Telecom Regulatory Authority of India (Trai) has issued draft guidelines for responsible AI use in telecom networks and services, The Economic Times reports. They would cover licensed telecom operators and other authorised entities that use or plan to use AI. These are proposals, not rules in force.
Under the draft, organisations would classify AI systems as high, medium or low risk, keep an inventory of active, inactive and decommissioned systems, and obtain approval from a designated officer before use. Systems would need to be listed in that inventory to be used.
The proposals also call for pre-deployment assessments of accuracy, robustness, explainability, fairness and resilience to model drift. Entities would set up a board-approved AI Governance and Safety Committee, review system performance at least annually, and report critical or high-severity incidents to Trai. Internal AI policies and amendments would be filed with the regulator within seven working days of taking effect.
Why it matters
The draft reaches beyond a checklist for model builders. It would put accountability on telecom entities for the outcomes of AI systems they use, including third-party products, and require processes for human oversight, consumer protection and handling reported harms.
That makes the proposals relevant to operators’ procurement and deployment choices, as well as their internal compliance teams. The key caveat is also the obvious one: these are draft guidelines, so their final requirements may change.
Our read
This is a substantial attempt to make AI governance part of telecom operations rather than a policy document left to gather dust. The inventory, approval and incident-reporting requirements give the proposal practical teeth. The next question is how Trai will define risk and turn broad assessment requirements into workable expectations for operators.
What to watch
- Whether Trai revises the draft before finalising it.
- How the regulator defines risk levels and incident-reporting deadlines.
- Whether the final framework sets out how compliance will be assessed.
Discussion spark: Should telecom operators face binding AI oversight rules for every system they use, including third-party tools, or should requirements depend more narrowly on the system’s risk?
Sources and evidence
- Source update (30 October 2026, 18:10 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.