
Meta has released an early technical overview of Scam Alert, an optional on-device tool for WhatsApp that flags likely scam messages without sending content to its servers. The August 12 release comes…
Meta has released an early technical overview of Scam Alert, an optional on-device tool for WhatsApp that flags likely scam messages without sending content to its servers. The August 12 release comes as India’s telecom regulator, TRAI, finalises rules that would bring WhatsApp, Signal and Telegram under network-level spam controls, requiring them to share spam reports with telecom operators’ Do Not Disturb registry. The Internet and Mobile Association of India (IAMAI) has opposed the draft as 'unconstitutional expropriation' of proprietary data.

Scam Alert runs a machine learning model on the user’s phone, drawing on scam conversations users have voluntarily reported. It triggers warnings for messages from non-contacts that match known scam patterns. Meta says it verifies model integrity through Cloudflare signatures and a public ledger, and publishes model weights for independent review. The Ministry of Electronics and Information Technology had earlier forced WhatsApp to halt its username feature over fraud concerns.

The usual binary is playing out again: TRAI’s draft rules are 'regulatory overreach' versus platforms 'enabling scams.' The truth is somewhere in the middle. WhatsApp’s Scam Alert shows real privacy-preserving tech exists, so the data-sharing demand in TRAI’s draft looks outdated. Yet same WhatsApp had to halt its username feature after government warnings on fraud. The real test will be whether TRAI’s final rules accommodate on-device detection like this, or forces a choice between encryption and spam control.
Source: medianama.com
This story was synthesised by AI from the source linked above.