Google Introduces AI-Driven Scam Detection for Android Messages and Calls

Google has unveiled a new security feature for Android devices aimed at combating fraud through the use of artificial intelligence (AI). This innovative system will monitor messages in Google Messages and phone calls, providing real-time alerts to users about potential scams. The technology is designed to operate locally on users' devices to ensure privacy.

The AI scam detection feature is an enhancement of previous safety measures introduced in October last year, which focused on protecting users from fraud and warning them about potentially harmful links. Recently, users have also noticed the Android System SafetyCore, a feature that alerts them to sensitive content in messages. Now, Google emphasizes its commitment to utilizing AI to identify fraudulent activities.

To develop this AI capability, Google has collaborated with financial institutions globally to better understand the most sophisticated and common scams affecting their clients. Often, these scams begin innocently but can lead to manipulation that causes victims to divulge personal information or transfer money.

Previously, the standard scam protection would alert users before conversations began, based on the phone number. However, many scammers disguise their identities or use numbers associated with trusted businesses to mislead victims. The new monitoring system in Google Messages will continuously scan exchanges, enabling Android to issue warnings during conversations if suspicious messages are detected.

Despite concerns regarding the surveillance of private conversations, Google assures users that the AI functionality operates solely on their devices. Only details about the scammer and their last known messages will be sent to Google and mobile service providers if a user decides to report a detected fraud attempt.

The scam detection capability will also extend to phone calls, with a feature already rolled out in the United States for certain Pixel smartphones by the end of 2024. During these calls, Android will alert users with sounds and vibrations if discussions involve payment-related topics. Similar to the message feature, Google emphasizes that these phone calls are not recorded or transmitted, although users must manually enable this feature.

The initial rollout of this AI security feature for Google Messages will target English-speaking countries, including the United States, Canada, and the United Kingdom. Plans for expansion to additional regions are in place, though timelines for German-speaking countries remain uncertain. In Germany, the telecommunications company Telekom has announced a SMS firewall set to launch in April, aimed at protecting users from malicious messages that could lead to malware installation or scams, but this service will focus solely on traditional SMS.