Google unveiled a suite of new AI-driven safety features in India aimed at strengthening user protection against online scams and fraud. Announced at the AI Impact Summit 2026, the updates are designed to ensure that “AI works harder for the defender than for the attacker,” the company said.
In a blog post, Google highlighted the increasing importance of building robust security directly into everyday devices and services. “Our devices are at the centre of our digital lives, so it’s crucial to integrate strong protections into the products we rely on daily,” the company noted.
Key Security Features Rolled Out in India
1) Real-Time Scam Detection on Pixel Devices
Google is introducing real-time scam call detection on Pixel smartphones, powered by its on-device Gemini Nano model. The feature analyses calls from unknown numbers in real time and warns users if the conversation resembles known scam patterns.
The tool is opt-in and must be enabled manually in the phone dialer settings. Google also confirmed that the feature operates fully on-device no audio is recorded or stored and includes an audible beep to notify both parties that detection is active.
2) Enhanced Protection for Financial Apps
To combat a surge in screen-sharing fraud, Google has partnered with Google Pay, PayTM and Navi.
Android devices (version 11 and above) will now display a prominent alert when a user opens one of these financial apps while screen-sharing with an unknown contact.
Users will also see a one-tap option to immediately end the call and stop screen-sharing to reduce the risk of financial scams.
3) Stronger System-Level Security with ePNV
Google has rolled out Enhanced Phone Number Verification (ePNV), a new Android security protocol that replaces traditional SMS one-time passwords with a secure, consent-based SIM verification process.
According to the company, this system raises baseline sign-in security across the ecosystem and reduces the effectiveness of OTP fraud.
4) Early Access to SynthID for Detecting AI-Generated Content
Google is expanding early access to SynthID, its AI watermarking and detection technology.
Through the SynthID Detector and API, users and developers can identify whether an image or text carries Google’s invisible watermark, signalling that it may have been generated by the company’s AI models.
Access is currently limited to approved early testers, with broader availability yet to be announced.
Source: livemint
