(Credit: LysenkoAlexander/Getty Images)
UPDATE 4/23: Google is now rolling out sensitive content warnings for Messages.
The feature will detect and blur images that may contain nudity in Google Messages, trigger a warning if you try to receive, send, or forward a nude image, and provide resources for you to get help when you receive this type of content, Google says.
Any detected nude images are stored on your device and won't be sent to Google. And it will probably make mistakes. "The sensitive content warnings feature isn’t perfect. It may occasionally detect images that don’t contain nudity or not detect images that contain nudity," Google warns.
For adults, this is turned off by default but can be turned on in Google Messages settings. Unsupervised teens can shut it off in settings. For supervised accounts, parents will have to turn it off in the settings of the Family Link app.
The enhanced scam detection mentioned below, meanwhile, rolled out last month.
Original Story 10/23/24:
Google Messages has announced five upcoming security features to make conversations on the Android app more secure and private.
First up is enhanced scam detection. The feature uses on-device machine learning, a subfield of AI, to detect scam texts including fraudulent package delivery texts and job scam texts. Once identified, it will automatically send them to the spam folder or alert the user. It's currently available for Google Messages beta users who have enabled spam protection.
Google Messages may also soon warn users about dangerous links they receive. It's already testing the feature in India, Thailand, Malaysia, and Singapore, with plans to expand it globally later this year. It cautions users when they receive links from unknown senders and can block messages with links if they're from suspicious senders.
Since spam texts often come from international numbers, Google Messages is also working on a feature that hides messages from unknown international numbers and sends them to the "Spam & blocked" folder automatically. Google will test the feature in Singapore later this year.
Another big addition to Google Messages is sensitive content warnings—something Apple and several social media platforms like Instagram already have. When enabled, the feature will blur messages containing nude imagery and present prompt that contains resources and options, but will ultimately still let recipients view the content. Google says that since the messages are end-to-end encrypted, it won't be able to access or see the nude images. The feature will roll out to devices running Android 9 or later in the coming months and will be enabled by default for users under 18.
Google is also developing a feature that helps you verify contacts and ensure you're texting the right person. To make this possible, Google is working on "a unified system for public key verification" across apps, which can be checked via QR codes or number comparison. The feature will work with Google Messages and a few other Android messaging apps, and will roll out sometime next year.
Google tested similar AI-powered scam detection in Gmail earlier this year, demonstrating at a conference how AI can identify phishing scams. It's also developed a feature for Android that uses Gemini AI to detect phone call scams in real-time.