In a first-of-its-kind initiative, Bharti Airtel has achieved what most consumers long considered impossible: considerably reducing spam and scam calls at the source.
Its newly deployed AI-based spam detection and blocking system has led to a nearly 70% reduction in cybercrime-related financial losses, according to data shared with the Indian Cyber Crime Coordination Centre (I4C) under the Ministry of Home Affairs.
According to the Indian telecom major, the technology works deep inside Airtel’s network, scanning over a trillion call and message records daily using 250 behavioural parameters, from call duration and frequency to device change patterns. In the first six months alone, the system flagged 26 billion spam calls and 1.4 billion malicious SMSes, shielding millions of users from potential financial fraud.
“This is not about call filtering at the handset. It’s about neutralising the threat before it ever reaches the user,” said a senior Airtel executive involved in the deployment.
Vi and BSNL Join the AI Defence Network
Not to be left behind, Vodafone Idea (Vi) has launched Vi Protect, an AI-led security suite for consumers and enterprises. Its system has already identified over 600 million spam and scam attempts, combining voice spam detection with link filtering and data-layer defences. Vi claims its solution is “up to 70% more accurate” and has sharply reduced false positives compared with legacy rule-based filters.
Meanwhile, BSNL, India’s public-sector telecom major, has introduced its AI/ML-based Anti-Spam and Anti-Smishing platform across its mobile network. It claims to detect 1.5 million scam attempts per day, blocks 35,000 fraudulent URLs monthly, and identifies 60,000 fake WhatsApp and mobile numbers each month — with a claimed 99% detection accuracy for malicious SMS links.
The TRAI push
The acceleration of such systems is no coincidence. The Telecom Regulatory Authority of India (TRAI) has mandated operators to curb unsolicited and fraudulent calls through network-level intelligence, rather than relying solely on user complaints. This has created a collaborative ecosystem where telcos share spam intelligence with cybercrime units and financial regulators — turning India’s telecom networks into active defence grids rather than passive conduits.
Challenges ahead
Despite impressive numbers, experts note that AI systems still face blind spots. Encrypted OTT calls (like WhatsApp or Signal), unregistered international VoIP numbers, and grey-market SIM farms remain difficult to track. False positives — legitimate numbers occasionally flagged as spam — are another challenge, though operators say accuracy improves continuously as AI models retrain on new data patterns.
Still, the results are undeniable. Airtel’s data-backed 68.7% drop in cybercrime losses shows that network AI is more than a buzzword — it’s saving real money and restoring digital trust. Vi’s Vi Protect and BSNL’s anti-smishing layer further demonstrate that the industry is moving in sync toward smarter, safer communication.
