Google and GovTech Singapore build on-device scam detection with Gemma
Google DeepMind and GovTech Singapore are building on-device fraud detection using Google's open model Gemma. DeepMind says more than half of reported crimes in Singapore involve sophisticated social engineering. Gemma is fine-tuned on local scam tactics so it can flag threats in real time without sending personal data off the phone.
Why it matters
It is a practical example of a small open model fine-tuned for one job and run on-device for privacy.
Who it's for
Developers working on safety or fraud, and anyone interested in on-device AI.
What it means for you
- Read DeepMind's post on the project.
- If you build fraud tools, note the pattern: small open model, local fine-tune, on-device inference.
- Look at Gemma's model cards to see which sizes run on phones.
Applies to: Project between Google DeepMind and GovTech Singapore.
Source: @GoogleDeepMind on X ↗
Written by AgentStackList from the original post. Checked Oct 17, 2026. Features, plans and availability can change, so confirm on the source before you rely on it.
gemmasecurityon-device
More Gemini tricks
- Nano Banana prompt pack: SQL-style "planet in a hand" city dioramas · Oct 17
- A developer reports high cache-miss rates on the Gemini API · Oct 17
- Turn a dense PDF into an interactive web app with Canvas · Oct 16
- Google AI Pro subscribers can now build with Argon in AI Studio · Oct 16
Agent Stack Weekly · free
Get the week's best AI tricks
Limit and pricing changes, new MCP servers, and setup fixes for Claude Code, Codex, Cursor, Copilot and Antigravity.
No spam. Unsubscribe anytime. Read past issues