Google DeepMind and Singapore Government collaborate to build fine-tuned AI model for scam detection
2026-10-26 · Google DeepMind
Google DeepMind and Singapore Government Collaborate to Build Fine-tuned AI Model for Scam Detection
Background: Singapore's Scam Epidemic
Scams including investment fraud, fake job offers, and impersonation of public officials constitute over half of Singapore's reported crimes. Since 2019, these scams have caused more than S$4 billion in losses. Critically, scam tactics have evolved from simple phishing to sophisticated social engineering, where attackers manipulate victims through convincing conversations before requesting money. Such scams now account for up to 80% of all scam-related losses.
Collaboration Overview
As part of the National AI Partnership with the Singapore Government, Google DeepMind has been working with the Government Technology Agency of Singapore (GovTech Singapore) to develop a privacy-preserving, on-device scam detection model. The model is built on Google's lightweight Gemma model, fine-tuned on local scam conversations covering impersonations of public institutions, banks, local figures, and trending regional schemes.
On-Device AI Protection Without Compromising Privacy
Generic models often struggle with local deception techniques. To address this, the Gemma model was fine-tuned on local scam conversation data, ensuring it can distinguish benign social chatter from genuine manipulation when deployed in citizen-facing applications.
Unlike traditional threat detection that relies on server-side cloud scanning, this model processes conversation data directly on-device. When integrated into applications, it enables faster analysis of conversation patterns to detect manipulative intent, helping citizens act before financial or credential exposure occurs.
The model's rollout will be carefully phased and developed in partnership with the Singapore Government. When deployed, the model will only access on-device conversations with the user's consent, in accordance with the application's terms of use. Since the model runs entirely on-device, neither the Singapore Government nor Google can access users' conversations through its deployment.
> Fighting scams shouldn't mean sacrificing personal privacy. Working with Google DeepMind, we're bringing advanced AI protection directly onto citizens' devices. This allows us to catch highly targeted, local scam attempts in real time while keeping personal communications private.
>
> — Mark Chen, Director, Trust and Safety, GovTech Singapore
Under the Hood: Local Context and Edge Efficiency
When the model detects manipulative conversational patterns, it triggers a real-time warning on the user's device, advising them to pause and exercise caution before sharing sensitive information or funds.
Running a complex scam classifier on modern smartphones presents significant challenges, particularly regarding battery efficiency. The team addressed this with a two-stage pipeline:
- Stage 1: A lightweight, efficient classifier screens inbound messages using minimal power, focused on catching potential threats.
- Stage 2: In-context analysis with the fine-tuned Gemma model, analyzing a larger conversation snapshot to confirm a threat or discard a false alarm.
Built on Strict Governance
Presented at Singapore International Cyber Week (SICW 2026), this collaboration provides a model for messaging platforms seeking to balance user protection with end-to-end encryption. Strict governance was maintained throughout development:
- Sovereign Sandbox: Fine-tuning took place within GovTech Singapore's managed and maintained GCP environment, keeping training data isolated.
- No Citizen Data: Training strictly excluded private or personally identifiable information, utilizing anonymised datasets derived from scam reports and intelligence feeds.
- Stress-Testing: The model was tested against synthetic dialogues and local corpora to cover diverse scam categories and evolving impersonation tactics.
The Future of On-Device Safety
This project demonstrates how AI can help counter criminality without compromising safety or privacy. It also illustrates how governments and model providers can collaborate to build technology for public good. With this lightweight, on-device model, the team has proven the capability for local, real-time scam detection while ensuring chat interactions remain secure. Google DeepMind looks forward to sharing more on its collaboration with the Singapore Government and exploring the potential for similar technology deployment in other countries.