Google and Singapore Launch AI Tools to Combat Scams

Google and the Government Technology Agency of Singapore are developing an on-device AI model for Android users to detect scams in real time. Announced during the Singapore International Cyber Week, the system analyzes conversational patterns on apps like WhatsApp and Telegram locally on the device to preserve privacy and bypass end-to-end encryption. The model is specifically fine-tuned to recognize Singaporean scam patterns, such as the impersonation of government officials and banks, and triggers warnings when threats are detected.

Google is simultaneously expanding its global Android scam defenses. A new opt-in enhanced protection setting in Google Messages uses cloud-based Gemini AI to identify slow-burn scam conversations from unrecognized senders. Additionally, the Gemini Nano AI model now inspects unfamiliar URLs for malicious signals on-device. To prevent fraud, Android will block apps from enabling call forwarding without notification and conceal sensitive data during screen-sharing in Chrome and Android WebView.

Google is also expanding its chat notification scam detection to a wider range of manufacturers, including Xiaomi, Oppo, and Motorola, to reach hundreds of millions of users by the end of the year. In Singapore, Minister Josephine Teo announced a S$30 million investment to extend the ASEAN-Singapore Cybersecurity Centre of Excellence and a new grant to help organizations migrate to quantum-resistant cryptography.