At Google‘s I/O developer conference yesterday, a breakthrough in phone security was unveiled: an AI-driven scam call detection feature aimed at shielding Android users from falling prey to phone scams.

This innovative technology, still in its testing phase, promises to alert users during calls if suspicious behavior or language indicative of a scam is detected. Developed by Google, the system employs Gemini Nano, a compact version of the company’s Gemini large language model optimized for Android devices. This model operates locally and offline, ensuring that user privacy remains intact while actively scanning conversations for potential fraudulent activity.

How does it work? Gemini Nano scrutinizes conversational cues that often accompany scam calls, such as unusual requests for sensitive information like passwords or PINs, demands for payment via gift cards, or urgent pleas for money transfers. If any of these red flags are detected, users will receive real-time alerts advising them to exercise caution or end the call.

Despite years of public awareness campaigns and educational resources on avoiding scams, the threat remains significant. According to a report by the Global Anti-Scam Alliance, a staggering 1 in 4 individuals worldwide fell victim to scams or identity theft in the last 12 months, collectively losing over $1 trillion.

Google has yet to disclose a specific release date for this feature, but users can anticipate the need to opt in to activate the protection. Compatibility, however, may initially pose a challenge for widespread adoption. Presently, Gemini Nano is only supported on select devices like the Google Pixel 8 Pro and Samsung Galaxy S24 series, as outlined on its developer support page.