← October 14, 2026 briefing · added in the 12:00 KST update

Google DeepMind and Singapore's government unveil AI tool using on-device Gemma to catch conversational scams in private chats

Announced October 14, 2026

Verified (fact)

Google DeepMind and GovTech Singapore, the Singapore government's technology agency, announced a new AI tool to protect citizens from scams. According to the Google DeepMind blog, the tool uses a lightweight Gemma model running on the device to filter out conversational scams in private chats. No data is sent to the cloud. The launch timeline and supported devices are not given in the collected excerpts.

Needs fact-check (claim)

The companies said the tool blocks scams in private chats with zero cloud data egress.

Why it matters (interpretation)

This is a case of a government putting a small open model on devices to build a public anti-scam service. It shows that security features can read conversational context without sending message content to servers. The setup could also be a useful reference for Korean institutions working on voice phishing and messenger scams.

Sources

More about Google