Key points
- Google is rolling out Enhanced Protection in Google Messages using Private AI Compute to analyse conversation context.
- On-device Gemini Nano models will assess unfamiliar links in real-time before users open them.
- Scam detection alerts are expanding from Pixel devices to a wider range of Android chat applications.
Google has announced a suite of new security protections for Android devices designed to combat mobile scams targeting calls, text messages and account verification codes. The update addresses the growing threat of criminals using spoofed numbers and AI-generated voices to trick users into surrendering financial data.
The company stated that its existing tools already identify more than five million suspected scam calls and texts each month. However, the new measures aim to prevent users from reaching the most dangerous stage of an attack by identifying suspicious behaviour before passwords or one-time passcodes are shared.
Context and background
Criminals are increasingly employing sophisticated tactics such as fake bank alerts, remote screen sharing and phishing pages to deceive victims. Google noted that standard URL filtering is often less effective against scammers who create new phishing domains that have not yet been added to threat blocklists.
To counter this, Android will utilise a combination of on-device artificial intelligence, cloud-based analysis and real-time checks. This approach allows the system to assess unfamiliar links by combining local analysis with real-time web signals before a user interacts with them.
Who is affected
The primary beneficiaries are Android users who receive unsolicited communications via text or chat applications. A key addition is an opt-in Enhanced Protection mode in Google Messages, which Google said will roll out in the coming weeks. This setting activates stronger protections for complex messaging scams involving unknown senders.
This mode uses Google’s Private AI Compute platform to allow Gemini models to review the context of a conversation. The aim is to spot slow-moving social engineering scams where criminals build trust over several messages before sending a fraudulent request.
Our take on the response
The expansion of scam alerts from Pixel devices to the wider Android ecosystem is a significant step for broader user protection. Scam Detection in chat notifications can now flag suspicious messages from supported chat apps directly in the notification panel. This provides an extra checkpoint, allowing users to see a warning before they even open the message.
What affected people can do
Users should review their Google Messages settings to enable Enhanced Protection once it becomes available. It is advisable to remain cautious of unsolicited messages from unknown numbers, even if they appear legitimate. Check for redesigned warning labels on links before opening them. Be wary of requests for one-time passcodes or remote screen sharing, as these are common indicators of fraud.
What to do and how to stay safe: Google
- Monitor notification panels for scam detection alerts in supported chat applications before opening messages.
- Verify the authenticity of any request for one-time passcodes or payment details through a separate, trusted channel.
- Avoid clicking on unfamiliar links in text messages until the on-device analysis has assessed their safety.
- Keep Android devices updated to ensure access to the latest on-device AI models and security patches.
General security guidance from the Breach Observer newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
When will the Enhanced Protection mode be available?
Google stated that the setting will roll out in the coming weeks.
How does the new system analyse message context?
It uses Google’s Private AI Compute platform, which allows Gemini models to review conversations in an encrypted environment.
Will these protections cover chat apps?
Yes, scam alerts are expanding to supported chat apps, flagging suspicious messages directly in the notification panel.