Google has announced the crackdown on artificial intelligence (AI) that support the functions of the scam to provide Android devices and their personal information.
“These features specifically focus on spoken scorches that can often appear initially harmless before turning into harmful situations,” Google – Note. “And even more phones that call use cheat methods to hide their real numbers and pretend they trust companies.”
The company said it cooperates with financial institutions to better understand the nature of the scammers who face customers, allowing it to develop AI models that can take suspicious samples and take real -time warnings during conversation without sacrificing users’ privacy.
These Models work completely on callPreventing users in case of probable scam. The users then have the opportunity to either reject or report and block the sender. Setup is included by default and spreads only to talking with phone numbers that are not on the device contacts.
The technical giant also emphasized that users’ conversations remain private and if they decided to report the chat as a spam, the sender’s details and recent messages with this sender are shared with Google and carriers.
This feature is launched in English first in the US, UK and Canada, and a wider expansion is scheduled for a later date.
In addition, a similar detection of scam for telephone calls is expanding in all English -speaking Pixel 9+ users. For the first time submitted In November 2024.
It is worth noting that the default feature is disabled to give users control, audio call is processed and never used during phone calls.
“If you turn on, the detection of the scam will be a signal at the start and during the call to inform the participants, the function is included,” Google said. “You can disconnect the scam at any time, during an individual call or for all future calls.”
Development comes a few weeks after Google has found that more than 1 billion Chrome users use advanced security security mode in the web browser.
“If you turn on increased protection, most of the additional protection you get from Adiv models and machine learning designed to detect dangerous URLL -url URLL – Note.
“Improved safe protection models can identify URLs designed to look like trusted domains. It also uses advanced AI and machine learning methods to protect you from dangerous downloads.”
