What Happened
Google announced a partnership with Indian carrier Airtel to address RCS spam in India. The integration deploys AI-powered spam filtering at the carrier infrastructure layer, rather than relying solely on device-side or application-layer filtering. TechCrunch reported the news on March 1, noting that India has faced disproportionately high RCS and SMS spam volumes for years, driven by both commercial bulk messaging and sophisticated scam operations.
Why It Matters
India is one of the largest mobile markets globally, with over a billion active mobile subscribers, and the scale of unwanted messaging is a genuine usability problem that has undermined consumer trust in SMS and RCS. Previous filtering approaches that operate on the device side or within messaging apps have limited effectiveness because they process messages after they've already been delivered to the network and often to the device.
Carrier-level filtering can intercept messages before they reach users' devices, which is a structurally better position for stopping spam at scale. Airtel has the network telemetry to train models on traffic patterns across all of its subscribers, not just individual users who have opted into filtering.
For Google, the partnership serves two goals: improving the RCS experience in a major growth market and demonstrating that RCS - Google's preferred successor to SMS - can match iMessage on practical quality. Apple's adoption of RCS in iOS 18 made the standard more relevant globally, but user experience depends on more than protocol support; it depends on whether the messages arriving are legitimate.
For Airtel, differentiation in the Indian carrier market on spam reduction is a real customer pain point that the company can lead on if the filtering works at scale.
Our Take
Carrier-level filtering is the right architectural layer for this problem, and Google has the machine learning capabilities to build it effectively. The partnership is a sensible infrastructure move. The broader question is whether improvements to RCS spam filtering translate into actual behavior change in a market where WhatsApp dominates personal communication and RCS is still establishing its position. Fixing spam is necessary but not sufficient to shift user habits.
The technical approach is sound. Training spam detection on carrier-level traffic patterns gives the model access to network metadata - sender reputation, volume patterns, known spam infrastructure - that device-side filters don't have. Whether the filtering is accurate enough to reduce false positives to an acceptable rate is the execution risk. Blocking legitimate transactional messages would erode user trust faster than the spam problem itself.