Android 17 is expected to begin rolling out later this month, but before that, Google has announced a batch of updates for the broader Android device ecosystem. As usual, some features are limited to specific devices, while others depend on Google's apps. However, if users don't mind relying on those apps, they can now gain automated protection against the growing threat of deepfake phone scams.

According to Google, impersonation fraud is one of the most common types of financial scams. The Federal Trade Commission tracked nearly $3 billion in losses from such scams in 2024, and advances in AI voice cloning technology are making these schemes easier to execute. Voice models have become so sophisticated that it can be difficult to identify a fake caller, even when the AI imitates someone you talk to daily.

How the new scam detection works

Google's solution expands on the system it debuted last month for verified financial calls. Now, a similar feature works with anyone in your contacts. Many effective deepfake scams involve spoofing a contact's phone number, making the call look legitimate when your phone lights up. Victims are then greeted by an accurate recreation of the person's voice telling an urgent story that requires cash.

Google's scam call detection will be available on all phones running Android 12 or higher, but it does require three Google apps: Phone by Google, Contacts, and Google Messages. Depending on your device, you may already have these preloaded—they come standard on Pixel and Motorola phones, and Samsung has now fully switched to Google Messages. Google claims Phone by Google is the most widely used dialer, though Samsung's own phone app remains dominant given Samsung's position as the largest Android OEM.

Once these three apps are installed, they will work together to verify calls that appear to come from a known contact. Scammers impersonating a contact use an online relay to spoof the number. When a call comes in, the caller's Google dialer app sends a confirmation signal that is absent in spoofed relay calls. If that signal is missing, your phone uses the Messages app to send an authenticated RCS ping (hence the Google Messages requirement) to the supposed caller. If their phone reports it is not placing the call, a pop-up alerts you that the person on the line may not be who they seem.

There is an important caveat: both parties must have the same three Google apps installed. If a caller uses the Samsung dialer or the OnePlus contacts app, Google's scam detection won't work. This limitation could reduce its effectiveness until more users adopt Google's communication apps.

As AI spoofing makes financial scams easier to execute, regulators and public safety organizations in some countries have advised Android users to avoid using their phones for important financial transactions. This poses a problem for Android and, by extension, Google. The latest anti-scam measure joins existing protections: Pixel phones can detect likely scam calls using on-device AI to identify suspicious caller behavior, and Google Messages already offers real-time scam identification.

More AI features: fashion and image search

No Android update would be complete without additional AI capabilities. The Find the Look feature in Circle to Search, initially launched on Pixel 10 and Galaxy S26 phones earlier this year, is now expanding to all devices running Android 14 or higher. Circle to Search already lets users search for images of anything on their screen. The new layer analyzes everything in an image, primarily to identify clothing. After circling an image, users can tap the Find the Look button to identify all parts of an outfit.

Google also announced that Google Photos will soon include an AI-assisted fashion engine. This feature catalogs the clothing you wear, creating a virtual wardrobe you can browse and organize from your phone. Users can even generate AI images of themselves wearing these outfits, though it's wise to see how clothes look in real life before leaving the house.

AirDrop support expands to more Android phones

In non-AI news, Google began supporting Apple AirDrop (via Nearby Share/Quick Share) on select devices earlier this year, but support was limited. Initially, only the last few generations of Pixel devices and Samsung's latest flagship phones supported the feature. Today, Google is enabling AirDrop on a wider range of Android phones:

• Samsung: Galaxy S25, S25+, S25 Ultra, S25 Edge, Galaxy Z Flip7, Z Fold7, Galaxy Z TriFold, Galaxy S24, Galaxy S24+, Galaxy S24 Ultra, Galaxy Z Flip6, Z Fold6
• OnePlus: OnePlus 15
• Xiaomi: Xiaomi 17T Pro (announced early)
• Vivo: Vivo X300 and X300 Pro
• HONOR: Magic V6

Nothing changes on the iPhone side. Sending files from Android still requires the iPhone to be set to accept AirDrop requests from 'anyone for 10 minutes.' Contact-based sharing is not supported when sending from Android, but the feature is becoming more widespread, making cross-platform file transfers easier.

Background and implications of the scam detection feature

The rise of deepfake voice technology has made phone scams more dangerous than ever. Scammers can now clone a voice from just a few seconds of audio—often captured from social media videos or voicemail greetings. They then spoof the victim's phone number using Voice over IP (VoIP) services that display any caller ID they choose. This combination of realistic voice and familiar number makes victims more likely to comply with urgent requests for money, such as paying fake bail, covering medical expenses, or helping a 'relative' in trouble.

The FTC estimates that in 2024, impersonation scams accounted for nearly $3 billion in losses, a figure that is expected to rise as AI tools become more accessible. Traditional call authentication methods, such as STIR/SHAKEN, verify the origin of calls but do not prevent number spoofing. Google's approach adds a secondary verification step that leverages the open RCS (Rich Communication Services) standard, which is already used for enhanced messaging. By having the called party's phone send an authenticated ping, the system can confirm whether the call originated from the actual device associated with the contact's number.

However, the requirement for both parties to use Google's specific apps limits adoption. Samsung users, for example, must switch to Google Messages and the Google dialer for the feature to work. While Samsung has already replaced its own Messages app with Google Messages on newer models, many users still have the old app. Similarly, OnePlus and other manufacturers often use their own dialers and contacts apps. Google has not announced plans to open this verification mechanism to third-party apps, which could slow the feature's impact.

The expansion of AI features in Circle to Search and Google Photos reflects Google's broader strategy to integrate AI into everyday tasks. Find the Look uses computer vision to recognize clothing items and provide shopping links or style ideas. The Photos wardrobe feature scans your existing photos to build a virtual closet, then uses generative AI to create images of you in different outfits. These tools are designed to increase engagement with Google's ecosystem and encourage users to store more data in the cloud.

AirDrop support is part of a larger effort to improve cross-platform interoperability. Google's Nearby Share was merged with Samsung's Quick Share to create a unified standard, and adding compatibility with Apple's AirDrop (via the new standard) reduces friction between Android and iPhone users. As more Android devices support this feature, the need for third-party file transfer apps will diminish. However, the iPhone's contact-based sharing restriction remains a hurdle, as users must temporarily set their device to accept from anyone.

Overall, these updates demonstrate Google's ongoing push to enhance security and convenience while driving adoption of its own apps and services. The scam detection feature, despite its limitations, represents a meaningful step forward in protecting users from AI-powered fraud. Combined with existing on-device protections, it adds a layer of verification that can alert users before they fall victim to a spoofed call. As deepfake technology continues to evolve, such proactive measures will become increasingly essential for mobile security.

Source: Ars Technica News