The Google Pixel 9 is receiving an upgrade for scam detection that you'll definitely want on your device.

      More than three months ago, Google began beta testing a new safety feature for Pixel phones that utilizes AI analysis to identify signs of fraud in voice calls. Today, Google has officially released the Scam Detection feature for calls, along with a similar system for screening messages.

      Every year, smartphone users lose millions of dollars to sophisticated scams around the globe. This issue is so widespread that the US Office of Inspector General and the Federal Trade Commission have issued guidelines on how to recognize and report such fraud.

      To identify potential scams during ongoing calls, Google is leveraging the natural language processing capabilities of its on-device Gemini Nano AI on Pixel phones. The AI listens to the conversation in real time and will display an alert if it detects any suspicious patterns.

      Users will receive notifications through audio cues, haptic feedback, and on-screen alerts indicating that the caller may be attempting to deceive them. Google has collaborated with financial institutions to analyze the most common tactics used by scammers.

      For instance, if the caller attempts to extract sensitive information like banking details, persuades users to download an app, offers free items, or tries to shift to another communication platform, the Scam Detection system will activate and raise an alert.

      Google has announced that Scam Detection for calls will be available to all Pixel 9 series users in the US, initially supporting the English language. The company tested this feature with other AI models on older Pixel phones, but found them to be less effective than the on-device Gemini Nano AI.

      Scam Detection will only be activated for numbers that are not saved in the user's contacts, and no audio recordings will be stored or sent to cloud servers. The implementation is entirely on-device. The feature is off by default and can be disabled during a call if desired.

      The goal is to proactively identify problematic patterns if initial defenses, such as spam call blocking, fail. “Traditional spam protections focus on safeguarding users before conversations begin and are less effective against newer tactics employed by scammers that become dangerous during calls and utilize social engineering methods,” states Google.

      Scam Detection in Messages

      The Google Messages app already features a strong set of spam reporting tools. It is now also incorporating Scam Detection, similar to what has been implemented for voice calls.

      Google indicates that the on-device AI will monitor SMS, MMS, and RCS conversations, alerting users when it identifies risky chat patterns.

      All message analyses occur on-device, and the scammer's details and messages are only shared with Google and carriers if users decide to report the sender as fraudulent.

      Unlike the calling feature, Scam Detection in Messages is enabled by default but can be turned off at any time. This functionality is being rolled out more broadly to English-speaking users in the US, UK, and Canada.