Google rolls out on-device scam detection and call-forwarding safeguards across Android
Backed by a Leviathan Security assessment, new AI-powered defenses analyze conversation patterns locally to flag fraud attempts.
Technology news
Google has begun rolling out an expanded suite of mobile fraud defenses across Android devices, introducing real-time scam detection, call-forwarding warnings, and screen-sharing data shielding. The release coincides with an independent evaluation by Leviathan Security Group comparing built-in anti-fraud protections across flagship smartphones from Google, Apple, Samsung, and Motorola.[5]
The headline feature brings on-device Scam Detection in Google Messages and compatible phone dialers. Powered by local machine learning via Android’s Private Compute Core, the system analyzes conversational patterns during active calls and text exchanges for indicators of social engineering, such as urgent wire transfer requests or fake courier alerts, without sending audio or message transcripts to cloud servers.[5]
Additional safeguards target account takeover tactics. A new carrier-level call forwarding monitor alerts users if unauthorized forwarding rules have been enabled that could intercept SMS-based two-factor authentication codes. Android also automatically hides sensitive one-time passcodes and banking interfaces during active screen-sharing sessions, mitigating tech-support scam exploits.[5]
Our take: on-device heuristic scanning provides a valuable line of defense against increasingly sophisticated conversational fraud, particularly for non-technical users. However, bad actors adapt quickly to avoid trigger phrases, and software warnings cannot prevent every user from voluntarily authorizing fraudulent transactions. These automated guards complement, but do not replace, fundamental security hygiene.