NEW DELHI– India is losing billions of rupees each year to online fraud, and Google is now rolling out new AI tools to protect smartphone users from scammers. Announced just before the upcoming AI Impact Summit in February 2026, the company is introducing on-device scam detection for Pixel 9 phones in India and new screen-sharing alerts for major financial apps.
These steps highlight Google’s focus on “safe and trusted AI”, but security specialists caution that these tools alone cannot stop the rising wave of complex cybercrime in the world’s largest democracy.
India’s online economy is growing at speed, with more than 1.2 billion mobile connections and UPI handling over 15 billion transactions every month. This success has also created fertile ground for fraud. Common scams include fake tech support calls, WhatsApp phishing, and “digital arrest” tricks where callers pose as police or officials and bully victims into transferring money.
According to TechCrunch , the Indian Cybercrime Coordination Centre, losses in 2024 crossed ₹20,000 crore (about $2.4 billion), with older adults and rural users facing the harshest impact.
In response, Google is using its Gemini AI models to turn devices into smarter gatekeepers that process threats on the phone itself, which helps reduce privacy risks.
Pixel 9 Gains Live AI Protection Against Scam Calls
The main highlight of the rollout is an on-device AI scam detection tool that will be available only on the Pixel 9 series in India through the November Pixel Drop update. Powered by Gemini Nano, Google’s compact AI model designed for on-device tasks, the feature listens to calls in real time and processes them locally so audio never leaves the phone.
Here is how the system behaves in practice. When a user receives a call from an unknown number, the AI listens for suspicious phrases, such as pressure to pay money quickly, callers pretending to be from a bank, or stiff, scripted “support” language.
If the AI thinks the call is likely fraudulent, a quiet banner appears on the screen with a message such as “Potential scam detected. Hang up?” and a one-tap button to end the call. The phone does not store recordings, and the detection runs fully offline, which helps in regions where mobile coverage is weak or unstable.
“Scammers move fast, but so does our AI,” said Sameer Samat, VP of Android and Google Play. “By processing detection on the device, we give users instant protection, especially when the stakes are high.” Internal trials in India, the UK, Australia, and Canada have shown encouraging results, with the system correctly identifying more than 70% of simulated scam calls.
For Pixel 9 users, a still small but growing slice of India’s premium phone market, that could mean fewer moments of panic when a strange caller demands an OTP.
The feature is not limited to calls. The same update extends protection to SMS and WhatsApp messages, scanning suspicious links and checking them against Google’s large threat index. In a country where roughly 80% of fraud begins with a misleading text or message, this early line of defence could prevent large losses.
Screen-Sharing Warnings for Banking and Payment Apps
Alongside call detection, Google is testing a feature aimed at one of India’s most aggressive fraud tactics: screen-sharing scams. Criminals often pose as customer service staff, then ask users to share their screen through Android’s built-in tools or remote access apps. Once inside, they guide the victim into banking or payment apps and drain their accounts.
To tackle this, Google has built a contextual warning system in partnership with Indian financial apps, including Google Pay, Navi, and Paytm. On phones running Android 11 or newer, if a user is on a call with someone and shares their screen, then opens one of these financial apps, a clear warning appears.
It reads: “You’re sharing your screen and opening a financial app. This could be a scam. End call and stop sharing?” A single tap lets the user cut the call and stop screen-sharing immediately.
The timing is important for India. UPI accounts for around half of all real-time payments worldwide, and data from the Reserve Bank of India shows screen-sharing fraud growing by about 300% year on year. “We are not just spotting threats; we are stepping in at the moment of risk,” Samat said. The pilot is expected to widen based on user feedback and links neatly into existing app permission systems, so users do not need another heavy security app that slows down their phones.
A Wider AI Safety Net and New Security Features
Beyond Pixel phones and pop-up alerts, Google is adding several upgrades across its ecosystem. Google Play Protect, which the company says has already blocked 115 million harmful app installs worldwide this year, will now focus more on scams that hit Indian users, such as fake trading apps and fraudulent investment platforms.
Early access to SynthID, Google’s watermarking system for AI-generated content, is also part of the package. SynthID tags audio or visual content created with AI so that platforms and users can detect deepfake voices or edited images used in scam videos.
There is also growing interest in ePNV, an electronic PIN verification method that could replace or support traditional OTPs. This system is intended to help reduce SIM swap fraud, where scammers take over a victim’s phone number to intercept OTPs and gain access to bank accounts. ePNV is still under trial, but banks and fintech companies are already watching it closely.
All these tools sit within Google’s larger “safe, trusted AI” message that the company plans to highlight at upcoming conferences. At the same time, India is pushing its own agenda around AI independence, including demands for data to stay within national borders and for better support for local languages and dialects in systems like Gemini.
The Obstacles: Limited Reach and Smarter Scammers
Despite the promise, Google’s security push has clear limits. On-device scam detection currently works only on Pixel phones, which represent a tiny share of India’s Android market. That leaves the vast majority of users on Samsung, Xiaomi, Realme, and budget devices without the same level of protection.
“Premium hardware gets premium protection; that is the Pixel tax,” said Priya Rao, a cybersecurity analyst at Bengaluru-based firm SecureNet. Bringing similar features to the wider Android Open Source Project could take many months, and handset makers would then need to adopt and distribute them.
Fraudsters are also learning fast. Cheap voice cloning tools and AI chat scripts are easy to find on underground forums. In controlled tests, Google’s detection accuracy sits around 80 to 85%. This is good enough for many day-to-day situations, but it may still miss some advanced or high-stakes attempts. Regulators such as the Reserve Bank of India have welcomed Google’s effort, yet they keep calling for shared action, including operators blocking spoofed caller IDs and stricter know-your-customer checks.
Education is another missing piece. Google is running awareness campaigns through YouTube Shorts and regional language content to explain how the alerts work. Many users still ignore warnings because they see too many of them. A LocalCircles survey in 2025 found that around 60% of Indians had dismissed scam alerts due to “alert fatigue”. For any AI tool to work, people need to trust it and understand what it is trying to say.
A Meaningful Step in the AI Security Battle
Google’s push in India signals a shift in how it applies AI: not just for flashy demos, but as a daily shield against fraud. In a country where scams can wipe out savings overnight and damage trust in online services, these features could recover some lost ground. For Pixel 9 owners, the phone effectively gains a smarter sense of danger, especially around calls and messages that feel off.
For everyone else, this rollout offers a glimpse of what AI-powered protection could look like across all Android devices in the future. As the AI Impact Summit approaches, many will watch to see if Google announces broader device support or deeper partnerships with banks and telecoms.
In the meantime, the advice is simple. If a caller claiming to be from “your bank” asks for an OTP or pressures for immediate action, and an alert on the screen suggests hanging up, it is usually wise to follow it. A single tap could be the difference between peace of mind and an emptied account.
