Android expands AI-powered protections against scam messages and phone fraud

09:10

By: Azzat Manal

Android expands AI-powered protections against scam messages and phone fraud

Google is expanding Android’s security features with new tools designed to help users identify fraudulent messages, avoid malicious links and protect sensitive personal information. The measures include an optional scam-detection feature powered by Gemini artificial intelligence, alongside additional safeguards for screen sharing, phone calls and suspicious activity.

The new option in Google Messages uses Google’s latest Gemini models to assess the content and apparent intent of incoming messages. When a conversation displays signs of a potential scam, the system can alert the recipient before the exchange develops further. One example presented by Google involves a stranger who pretends to have sent a photograph to the wrong number, then attempts to keep the conversation going by asking personal questions. Such exchanges can be used to establish trust before a fraud attempt.

Google says the system is designed to process messages within its Private AI Compute environment, which uses protected computing infrastructure to handle sensitive information. The company says this approach is intended to preserve user privacy while enabling advanced AI-based analysis. Android will also use Gemini Nano to assess links directly on compatible devices, helping users identify potentially dangerous destinations before opening them.

The new protections are expected to reach selected devices, including Google’s Pixel 10 series and newer models, Samsung’s Galaxy S26 range, and the Galaxy Z Flip 8 and Galaxy Z Fold 8, alongside other compatible smartphones. Google is also redesigning scam warnings to make them clearer and give users more time to assess links received from numbers that are not saved in their contacts.

The company is extending scam alerts beyond the messaging app itself. Following the introduction of scam detection in notifications on Pixel devices, Google is working to bring similar warnings to a wider range of Android phones. The feature is designed to notify users about suspected fraud directly through message notifications, allowing them to receive a warning without opening the conversation. Device manufacturers including Honor, Motorola, Oppo, OnePlus, Realme, Vivo and Xiaomi are expected to introduce the capability on compatible products.

Privacy protections are also being expanded to cover screen sharing. Starting in late October, Android 15 and later versions are expected to receive automatic safeguards that conceal sensitive information, such as passwords, banking details and credit card security codes, when users share screens displaying content from Chrome or Android WebView. The measure could reduce the risk of accidentally exposing confidential information during technical support sessions, video calls or other situations in which a phone’s screen is visible to someone else.

Google is also planning a call-forwarding protection feature for early December on Pixel phones and selected Android devices. The tool is intended to prevent applications from secretly enabling call forwarding in the background and warn users when they are prompted to enter a mobile network code, commonly known as an MMI code, that could activate the service. Criminals may exploit such codes to redirect incoming calls to another number, potentially exposing sensitive communications.

Separately, Google reported progress in tackling fraudulent calls in Brazil, where Android has been working with banks to verify whether incoming calls genuinely originate from the institutions they claim to represent. Calls that fail the verification process can be terminated, with users receiving an alert. Google said the system had blocked more than three million fraudulent calls since its launch in June. The company has not announced when this particular protection might become available in the United States, but says it intends to work with additional banking partners to expand its reach.

Together, the measures reflect Google’s