ResearchPublished All news from that day
DeepMind and GovTech Singapore build an on-device scam model, with a phased rollout planned.
Google DeepMind says the Gemma-based model is fine-tuned on local scam conversations and will run entirely on the device once deployed.
Check our sources · 6 facts from 1 sourceGoogle DeepMind said on 14 October 2026 that it worked with the Government Technology Agency of Singapore on a fine-tuned, on-device scam detection model.
It is a Gemma model fine-tuned on local scam conversations, and once deployed it will run entirely on the device, so neither the Singapore Government nor Google will be able to access users' conversations through it.
A lightweight classifier screens inbound messages first, and the fine-tuned model then analyses a larger snapshot of the conversation to confirm a threat. Fine-tuning took place in GovTech Singapore's managed GCP environment on anonymised datasets from scam reports and intelligence feeds.
DeepMind says rollout inside applications will be phased and developed with the Singapore Government.
Underlined sentences link to the facts we checked. Dotted words open a short explainer.
Fact 1 of 6 · from Google DeepMind, 14 Oct 2026
we've fine-tuned a privacy-preserving, on-device scam detection model
Fact 6 of 6 · from Google DeepMind, 14 Oct 2026
The roll out of the model within applications will be carefully phased and developed in partnership with the Singapore Government.
Your reaction
We count reactions per story and day, never who reacted. The counts help us choose what goes in the monthly issue. If you are signed in, your own page shows yours too.
We count reactions per story and day, never who reacted. They help choose what goes in the monthly issue.
Check our sources
Every sentence above is checked against this source.
1 Google DeepMind and Singapore Government collaborate to build fine-tuned AI model for scam detection
Open the source archived copy-
Google DeepMind published a post on 14 October 2026 saying it worked with the Government Technology Agency of Singapore on a fine-tuned, privacy-preserving, on-device scam detection model.
citewe've fine-tuned a privacy-preserving, on-device scam detection model
-
Google DeepMind says the model is Google's lightweight Gemma model fine-tuned on local scam conversations.
citeGoogle’s lightweight Gemma model was fine-tuned on local scam conversations
-
Google DeepMind says that once deployed, the model will run entirely on-device, so neither the Singapore Government nor Google will be able to access users' conversations through it; it can only read on-device conversations with the user's consent. It is not deployed yet.
citeWhen deployed, the model will only be able to access on-device conversations with the user's consent, in accordance with the terms of use of that application. As the model runs entirely on-device, neither the Singapore Government nor Google, will be able to access users' conversations through its deployment.
-
Google DeepMind says the model uses a two-stage pipeline in which a lightweight classifier screens inbound messages with minimal power, and the fine-tuned Gemma model then analyses a larger snapshot of the conversation to confirm a threat or discard a false alarm.
citeStage 1: A lightweight, efficient classifier screens inbound messages, using minimal power, focused on catching potential threats. Stage 2: In-context analysis with our fine-tuned Gemma model, where we analyze a larger snapshot of the conversation to confirm a threat or discard a false alarm.
-
Google DeepMind says fine-tuning took place within GovTech Singapore's managed Google Cloud environment, and training excluded private or personally identifiable information, using anonymised datasets derived from scam reports and intelligence feeds.
citeFine-tuning took place within GovTech Singapore's managed and maintained GCP environment ... Training strictly excluded private or personally identifiable information, utilising anonymised datasets derived from scam reports and intelligence feeds.
-
Google DeepMind says the rollout of the model within applications will be carefully phased and developed in partnership with the Singapore Government.
citeThe roll out of the model within applications will be carefully phased and developed in partnership with the Singapore Government.
Every sentence above is checked against this source. Each fact has its own address you can cite.
Source 1 · Google DeepMind · 14 Oct 2026
Google DeepMind and Singapore Government collaborate to build fine-tuned AI model for scam detection
Google DeepMind published a post on 14 October 2026 saying it worked with the Government Technology Agency of Singapore on a fine-tuned, privacy-preserving, on-device scam detection model.
we've fine-tuned a privacy-preserving, on-device scam detection model
Google DeepMind says the model is Google's lightweight Gemma model fine-tuned on local scam conversations.
Google’s lightweight Gemma model was fine-tuned on local scam conversations
Google DeepMind says that once deployed, the model will run entirely on-device, so neither the Singapore Government nor Google will be able to access users' conversations through it; it can only read on-device conversations with the user's consent. It is not deployed yet.
When deployed, the model will only be able to access on-device conversations with the user's consent, in accordance with the terms of use of that application. As the model runs entirely on-device, neither the Singapore Government nor Google, will be able to access users' conversations through its deployment.
Show all 6 factsShow fewer facts
Google DeepMind says the model uses a two-stage pipeline in which a lightweight classifier screens inbound messages with minimal power, and the fine-tuned Gemma model then analyses a larger snapshot of the conversation to confirm a threat or discard a false alarm.
Stage 1: A lightweight, efficient classifier screens inbound messages, using minimal power, focused on catching potential threats. Stage 2: In-context analysis with our fine-tuned Gemma model, where we analyze a larger snapshot of the conversation to confirm a threat or discard a false alarm.
Google DeepMind says fine-tuning took place within GovTech Singapore's managed Google Cloud environment, and training excluded private or personally identifiable information, using anonymised datasets derived from scam reports and intelligence feeds.
Fine-tuning took place within GovTech Singapore's managed and maintained GCP environment ... Training strictly excluded private or personally identifiable information, utilising anonymised datasets derived from scam reports and intelligence feeds.
Google DeepMind says the rollout of the model within applications will be carefully phased and developed in partnership with the Singapore Government.
The roll out of the model within applications will be carefully phased and developed in partnership with the Singapore Government.
Topics
The morning email
On the mornings we publish: the three top stories and up to four short ones. Free.
We email you a link to confirm. An issue may include one sponsor, always labelled Sponsored · Advertisement. Our emails count opens and clicks, not who made them. Unsubscribe in one click. What we keep