TRAI directs telcos to share AI-based spam data on DLT platform within hours – Indian Infrastructure
The Telecom Regulatory Authority of India (TRAI) has directed all mobile operators to share data generated by their artificial intelligence (AI)-based anti-spam systems on the common blockchain-based distributed ledger technology (DLT) platform within hours and initiate action against suspected spammers, even in the absence of consumer complaints.
The directions, issued on February 27, 2026, require operators to comply within 30 days.
While TRAI has not explicitly mandated the blocking of suspected spam numbers, after operators raised concerns over draft proposals, it has fixed accountability on both the originating and terminating telecom operators to coordinate and take appropriate action.
Under the framework, the terminating operator must use its AI/ machine learning (ML)-based spam detection system to identify and flag a calling line identification (CLI) or mobile number as “suspected spam” based on defined behavioural parameters. The flagged CLI must then be shared with the originating operator via the DLT platform immediately, and in any case within two hours of identification.
The originating operator is required to inform the sender that the number has been flagged and verify the sender’s know-your-customer credentials. The shared data on the DLT platform will enable all operators to identify telecom resources linked to the sender and assess whether additional numbers associated with the same entity have been flagged as potential spam.
If five or more CLIs associated with the same sender are identified as potential spam within a 10-day period, operators are required to initiate action against the sender.
Telecom operators had earlier expressed reservations about acting solely on AI-generated flags, arguing that such systems are not fully accurate and may wrongly identify legitimate numbers as spam. They had favoured continuing the existing mechanism, where action is triggered by consumer complaints. Operators have also pointed out that their AI systems differ in technical parameters, which may pose challenges in harmonising shared data.
TRAI, however, has maintained that merely alerting subscribers without backend enforcement does not provide sufficient deterrence. The regulator noted that nearly 85 per cent of spam complaints relate to unregistered telemarketers and said effective containment requires calibrated use of AI/ML-based network intelligence deployed by telecom service providers.
