Google has announced an increased rollout of new AI-powered scam detection features on Android to help protect users from increasingly sophisticated phone and text social engineering scams.
The tech giant says these features come in response to AI-assisted scam campaigns that, according to statistics from the Global Anti-Scam Alliance, have defrauded people of over $1 trillion.
“Today, we’re launching two new industry-leading AI-powered scam detection features for calls and text messages, designed to protect users from increasingly complex and damaging scams,” reads Google’s announcement.
“These features specifically target conversational scams, which can often appear initially harmless before evolving into harmful situations.”
Google warns that traditional spam protections designed to block these attempts are ineffective at stopping sophisticated scams that manifest after the victim has already started a conversation with the fraudsters.
To identify and stop these scams, Google has partnered with banks and other key entities to understand the scams consumers are facing.
“To enhance our detection capabilities, we partnered with financial institutions around the world to better understand the latest advanced and most common scams their customers are facing. For example, users are experiencing more conversational text scams that begin innocently, but gradually manipulate victims into sharing sensitive data, handing over funds, or switching to other messaging apps. And more phone calling scammers are using spoofing techniques to hide their real numbers and pretend to be trusted companies.
Traditional spam protections are focused on protecting users before the conversation starts, and are less effective against these latest tactics from scammers that turn dangerous mid-conversation and use social engineering techniques. To better protect users, we invested in new, intelligent AI models capable of detecting suspicious patterns and delivering real-time warnings over the course of a conversation, all while prioritizing user privacy. “
The new scam-protection features
The first set of security features is AI-powered enhancements of existing spam protection mechanisms in Google Messages, Android’s default SMS and RCS messaging app.
Google says ‘Scam Detection’ for Messages will now detect a broader range of scam attempts, including job and item delivery lures, warning users and giving the option to report and block the sender.
The system, which is enabled by default, will only be activated when receiving messages from unknown senders whose numbers aren’t in the user’s contacts list.
The initial rollout will cover Android users in the U.S., the U.K., and Canada.
A similar system will be introduced for calls, where Google’s AI tool Gemini Nano will analyze conversations in real time to identify common fraud attempts, such as detecting signs of people asking for payments via gift card codes. The feature would alert the user via audio and haptic notifications when a scam is detected.
Source: Google
This feature was first announced in November 2024 and has been tested on a small set of Pixel 6+ devices, but it is now being rolled out to all English-speaking Pixel 9 users in the U.S..
Pixel 9 users will get the feature through Gemini Nano, while Pixel 6 to 8 series devices enrolled in Phone by Google beta will be restricted to less powerful Google AI models.
Both the Scam Detection for Messages and voice security features analyze the conversation data on the device without sending any sensitive information to Google, so the new features respect users’ privacy.
Also, a characteristic “beep” will play at the start of the conversation, audible by both parties, to notify participants that AI-based analysis is taking place.
Still, Google decided to keep Scam Detection for calls disabled by default to relieve privacy concerns.
Users can enable it by opening the Phone app, tapping the three-dot menu in the top right corner, selecting Settings → Scam Detection (if available), and toggling the feature to the ON position.
When Scam Detection is active, users are still given the option to turn it off during an individual call.