Google Messages to Blur Inappropriate Pics

Google Messages can now warn users before they open an inappropriate photo. Privacy experts say the technology behind the tool appears to be safe.

Officially the feature is for "Sensitive Content Warnings" though it appears to actually be all about photos of undressed bodies or body parts. The main focus is dealing with unsolicited images, though it could also prevent unintentionally opening of an provocative image in public when a phone screen could be seen by others.

When enabled, the feature will initially blur the image. It will then give the user options including viewing the image and blocking the sender's number.

By default the feature will be switched off for adult users and enabled for under-18s, whose accounts are set up through the Family Link tool. Those 13-17 years olds set up with an "Unsupervised" account will be able to switch the feature off.

Those on a "Supervised" accounts can't directly switch if off themselves, but the parents can do so through Family Link, which would presumably follow quite the family dinner chat. (Source: techradar.com)

Privacy Protected

Privacy advocates were naturally concerned about how Google is identifying such images and what scanning and storage is involved. However, it appears there's no significant risk here.

The assessment of the image is done by a technology called SafetyCore that runs entirely on the phone and doesn't store or send any information about images to Google or any other organization. It's a similar technology to that used to assess messages for scam or malware. (Source: forbes.com)

Senders Asked to Think Twice

The feature also works for people sending or sharing messages with such photographs, though in this case it's not a case of blurring the image but rather giving a warning about potential harm from such images. The user can then confirm whether they really do want to send the image.

That seems somewhat pointless as it's unlikely anyone intending to send such images in the first place is going to switch on an optional feature that warns them against doing so.

What's Your Opinion?

Is this a sensible feature? Is it a technical 'answer' to a social 'problem'? Do you trust the scanning to be as secure as it appears?