Theft Lock technology identifies a specific sequence of events to determine that a theft is in progress and automatically locks down the phone, preventing attackers from accessing the victim’s apps and financial information. 

"We invest a lot in security and privacy, it’s core to what we do,” said Android’s VP of Engineering Dave Burke, speaking at a roundtable on the sidelines of the annual Google I/O event where the new feature was announced. “We want to push the technology forward to be better at fighting theft and fraud.” 

The problem of phone theft in Brazil helped drive the development, explained Android Ecosystem President Sameer Samat, also present at the roundtable. 

“Brazil was a big influence in the creation of the Theft Lock features,” he said. “I went to Brazil and talked to consumers and government and they said this was a big issue, a big problem.” 

The team shared statistics showing that a phone is stolen every five minutes in São Paulo, and every six minutes in London. In Brazil, Burke said, the typical scenario involves someone on a bike or motorbike who grabs the phone from the user’s hand in a drive-by snatching. 

That’s the sequence that the AI is trained to detect: an unlocked phone that’s in use, experiences a sudden impact and then rapid movement (running or on a vehicle) triggers the lockdown. 

Burke said that testing has shown very few false positives, and the feature will be available across the Android ecosystem, not just on the upcoming Android 15 release. 

But it’s in the next iteration of the operating system that additional anti-theft and anti-fraud features will appear. One is the creation of a virtual private space within the phone (Private Space) that lets users keep sensitive apps in a space that only the legitimate user can access. 

There will also be measures to combat the growing number of scams that target users, such as attacks where fraudsters pretend to be the user’s bank, send a phishing link and are able to see the person’s screen. Android will hide one-time passwords and codes sent via text message to prevent these attackers from seeing them during screen sharing. 

The keynote at Google I/O also showed off a future feature made possible by AI: the ability for the system to detect, mid-call, that the user is being targeted by a scam and prompt them to hang up. 

The 2024 edition of the annual Google I/O event was all about the company’s progress in Artificial Intelligence, which is being integrated into everything it does — from search on its search engine to documents in Google Workspace, Gmail and the Android mobile operating system. 

Read Also: Apple and Google will alert users about possible tracking (Portuguese version)