The Telecom Regulatory Authority of India (TRAI) has amended its commercial communications rules to mandate telecom operators to take action against suspected spammers flagged by their AI-based caller ID systems, even without requiring consumers to file complaints. According to the latest notification published on Monday, every telecom operator receiving an incoming call flagged as 'suspected spam' by its artificial intelligence or machine learning enabled system will have to share details of such calls with telecom operators from whose network the call has originated within two hours of such identification. The Telecom Commercial Communications Customer Preference (Third Amendment) Regulations, 2026 (05 of 2026) introduce a formal definition of 'spam' and 'spammer' and expand the framework to cover application-to-person (A2P) calls. Under the new rules, AI systems deployed by telecom operators flagged about 771 crore calls and 48 crore SMS as suspected spam in April 2026, as cited by TRAI data. The regulator's analysis shows that AI systems flagged on average more than 25 crore calls and 1.6 crore SMS per day as suspected spam during April 2026, with these systems largely used to alert subscribers rather than effectively integrated into enforcement against UCC senders.
The amended framework introduces a three-complaint threshold where regulatory intervention can be initiated when three or more unique recipients complain against a sender within 10 days, and any CLI allotted to that sender has also been identified or maintained as a 'suspected high probability UCC CLI' by operators' AI systems during the same period. According to TRAI's analysis, applying the three-complaint threshold would have brought 28,135 complaints and 6,762 senders within its ambit, substantially expanding the pool of cases potentially subject to investigation. The regulations seek to strengthen action against unsolicited commercial communications by combining consumer complaints with intelligence generated by AI and machine-learning based spam-detection systems, with TRAI clarifying that AI flagging alone will not result in punitive action - AI-generated intelligence while a strong indicator of possible UCC activity, cannot be treated as conclusive evidence because false positives cannot be entirely ruled out. Under the new rules, originating access providers (OAP) shall, within one business day of the receipt of the flagged calls, identify unique KYC identifiers of the sender associated with such calls and share those details with other telecom operators. As per the new regulations, all operators then have a 24-hour window to review their internal systems for any other connections or resources linked to that same sender over the preceding 10 days.
For the first violation triggered using AI systems, outgoing services of all telecom resources, including allotted resources to the sender, shall be barred by all telecom operators for a period of 15 days, regardless of whether those telecom resources were actually used in making such communications. For subsequent instances of violations, all telecom resources, including PRI/SIP trunks, SIMs etc. of the Sender shall be disconnected by all the Access Providers for one year, irrespective of whether those telecom resources were actually used. The violator will still be provided with one telephone number with outgoing services barred, allowing connection only with notified emergency outgoing services. In cases involving an unregistered telemarketer, a first established violation can result in outgoing services being barred for 15 days, while for subsequent violations, all telecom resources can be disconnected across access providers for one year, and the sender can also be blacklisted for one year, during which access providers cannot issue fresh telecom resources. The telecom operators or the OAP will have to put the sender under the blacklist category during the period of one year and no new telecom resources shall be provided by any telecom operator to the violator.
Subscribers opting to block telemarketing calls from all categories will need to text BLOCK PROMO code to 1909, as reported by Upstox News Desk. The 'BLOCK PROMO option shall block only promotional types of Commercial Communications for all categories of content, mode, time band and day types but shall not block transaction and service type of Commercial Communications and Government Communications'. The amended rules also clarify that messages containing OTPs for services such as app login, password reset and two-factor authentication will fall under service messages. Consumers will continue to be able to opt out of promotional commercial communications through 1909, either across all promotional categories or selectively for sectors such as banking and financial products, real estate, education, health, consumer goods and automobiles, communication and entertainment, tourism and leisure, and food and beverages. All preferences registered under the FULLY BLOCK category under present rules will be migrated to BLOCK PROMO category within fifteen days of the notification.
The amended rules will be implemented in phases of 30-90 days from the date of notification, with most provisions coming into force after 30 days, specified provisions after 60 days, and amendments listed in the annexure after 90 days from publication in the Official Gazette. According to the notification, all telecom operators within one day will need to check within their system if there are other connections within their network linked to the suspected spammer in the last 10 days. The regulations provide that if it is found that five or more CLIs (called line identification) of the Sender have been so identified within the last ten days, all the OAPs concerned shall take action against the Sender. The amendments also formally bring automated voice communications within the framework by defining an Application-to-Person (A2P) call as one initiated by an application, software system or automated platform without direct human dialling, with senders using A2P calls required to declare in advance the use of such calls and details such as CLI ranges to be used. Under the amended rule, Terminating Access Provider (TAP) can charge the OAP up to ₹0.05 (five paisa only) per minute for A2P calls, with the notification also barring all call management applications or similar services like Truecaller from tagging, blocking or filtering incoming calls originating from any number series designated by the authority like those starting with 160 or the central government for commercial and government communication. The remaining provisions including a blockchain-based appeal mechanism, AI-driven data-complaint corroboration tools and automatic service suspensions for senders misusing content templates will take effect after 90 days.
