Skip to content
Google DeepMind and GovTech Launch On-Device Scam Detection Model

Google DeepMind and GovTech Launch On-Device Scam Detection Model

First seen 14 Oct 2026, 14:38 UTC • •
API

Article Content

Browse articles
ThreatCluster AI
ThreatCluster •October 14, 2026 at 17:35 UTC
  • •New on-device scam detection model launched by Google DeepMind and GovTech.
  • •Scams in Singapore have caused over S$4 billion in losses since 2019.
  • •Model processes data locally to enhance privacy and efficiency.

On October 14, 2026, Google DeepMind and the Government Technology Agency of Singapore (GovTech) announced a new on-device scam detection model aimed at protecting smartphone users from manipulative conversations. This model addresses the rising issue of scams in Singapore, which have led to over S$4 billion in losses since 2019, with complex social engineering tactics now accounting for up to 80% of these losses. The model, named Gemma, is optimized for local scam conversations and processes data directly on users' devices, ensuring privacy by avoiding cloud data transmission. It features a two-stage pipeline that efficiently screens messages and confirms potential threats. The initiative was unveiled during Singapore International Cyber Week 2026, highlighting the collaboration under the National AI Partnership with the Singapore Government.

Start a free Starter trial for enhanced analysis

Ask AI about this cluster

Updated just now How this analysis works

Timeline

2026-10-14
On-device scam detection model announced
Google DeepMind and GovTech unveiled the Gemma model to combat rising scams in Singapore.
Unite.Ai
2026-10-14
Collaboration presented at Singapore International Cyber Week
The collaboration was highlighted during the event, emphasizing the importance of AI in cybersecurity.
deepmind.google

More articles in this cluster (2)

Following this threat?

Track Google Cloud Platform in your own feed — you're alerted when they show up in new reporting, leak sites or exploitation.

Free account · no card needed

Extracted Entities (4)

1 / 2

Common questions

How does the on-device model work?
The model processes conversation data locally on smartphones, allowing for real-time threat detection without sending data to the cloud.
What types of scams does it target?
It targets various scams, including impersonations of public institutions and complex social engineering tactics.
Is user privacy compromised?
No, the model is designed to protect user privacy by analyzing data locally without transmitting it to external servers.