TRAI tightens anti-spam rules, brings AI detection and robocalls under new framework

The Telecom Regulatory Authority of India (TRAI) on Friday strengthened its rules against unsolicited commercial communications, placing artificial intelligence-based spam detection inside the formal regulatory framework and giving consumers a clearer route to appeal if their complaints are poorly handled.

The regulator notified the Telecom Commercial Communications Customer Preference (Third Amendment) Regulations, 2026, amending the 2018 code that governs promotional calls and messages.

These changes follow a draft released in March and several months of talks with operators, telemarketers, and consumer groups.

Officials say the goal is to shift from slow, complaint-based enforcement to quicker, intelligence-driven action as robocalls and automated calls have increased. The new rules require telecom providers to use AI and machine learning to identify numbers likely used for spam.

Operators will share suspected numbers using the current distributed ledger platform. If five or more numbers linked to the same sender are flagged within ten days, providers must investigate and take steps such as new know-your-customer checks, physical verification, blocking outgoing calls, and disconnecting services for repeated misuse.

The threshold for action on complaints has also been lowered. Now, action can start after three unique complaints in ten days if the operator’s AI system also flags the sender’s number. Previously, five complaints were needed, no matter what the system detected. TRAI says this change helps match consumer reports with network data and cuts down the time serial offenders have.

The amendment now clearly defines Application-to-Person calls as those started by software or automated systems, not by a person dialling each number. This includes auto-dialers, robocalls, and calls with pre-recorded or artificial voices. Anyone using these systems must tell their operator in advance and share the numbers involved. Calls made without this declaration will count as unsolicited commercial communication.

A termination charge of up to five paise per minute now applies to A2P calls, except for calls on certain approved commercial number series and those authorised by the Authority. Special number ranges used for regulated commercial messages, like 140xx, 1600xx, and 1601xx, will not be marked as suspected spam for recipients.

Call-management apps cannot block or label these number series as junk by default. User spam reports must also be sent to the operators’ ledger system, not just kept in third-party apps. If consumers are unhappy with how their spam complaint is handled, they can now appeal within 15 days using the same channels they used to file the original complaint, such as operator apps, portals, the DND app, or 1909.

A senior employee of the service provider, acting as the appellate authority, must decide on appeals within 15 days. TRAI Chairman Anil Kumar Lahoti said operators will need to re-verify high-risk numbers and disconnect those found to be used for spam.

The regulator has also shortened the time businesses can contact a customer after an inquiry to seven days, unless they get new consent. The problem is widespread.

In the first quarter of this financial year, operator AI systems flagged about 22.99 billion suspected spam calls out of around 721.79 billion incoming calls. Officials say sending alerts alone has not stopped senders, so investigations and penalties are now needed.

Industry groups mostly support faster detection but warn that false positives could affect legitimate business calls.

TRAI says that behavioural flags must be confirmed and that approved commercial traffic will stay protected. Now, access providers must update their detection systems, sharing methods, and customer appeal processes to match the new rules.