Google DeepMind and Singapore Deploy On-Device Fraud Detection Using Gemma AI Model

Google DeepMind and Singapore's Government Technology Agency (GovTech) announced a partnership to deploy on-device scam detection using a fine-tuned Gemma language model, protecting citizens from social engineering attacks while preserving message privacy. Scams account for more than half of reported crimes in Singapore, with S$4 billion in cumulative losses since 2019. The new system processes conversation data directly on users' devices rather than sending it to cloud servers, maintaining end-to-end encryption on messaging platforms while analyzing chat patterns for manipulative intent.