Google Begins Scanning All Your Photos—Here’s How to Stop It with One Click
A few weeks ago, Apple found itself in a bit of a mess when users discovered that their photos were being scanned by Apple’s technology to identify landmarks. The problem wasn’t really with the technology itself; it was more about the fact that users weren’t informed about this process, which upset many people, especially security experts. Now, Google is experiencing a similar situation, and once again, it’s about the lack of transparency.
Apple uses a feature called Enhanced Visual Search, which sends parts of users’ images to the cloud so they can be compared to a global database of landmarks. While this system is designed to protect privacy, many, like crypto expert Matthew Green, expressed frustration when they learned about this feature just days before New Year’s, only to find it was already active on their phones.
Google’s recent controversy centers around its new SafetyCore feature, part of an Android system update. SafetyCore enables image scanning directly on devices, focusing on blurring or flagging sensitive content. This is meant to be very private since it works entirely on the device, unlike some services that operate in the cloud.
However, the main concern arises when features like SafetyCore are activated on users’ phones without any prior notice. When users are informed after the fact, it leaves them skeptical, similar to how many felt about Apple’s Enhanced Visual Search.
Although I have previously discussed SafetyCore and noted that shifting security scanning from Google’s servers to users’ phones could be beneficial, the underlying issue of transparency still remains. GrapheneOS, a security-focused Android developer, offers some reassurance that SafetyCore doesn’t report scanned data to Google. It uses local machine learning to classify content, allowing apps to check certain content without sending it elsewhere.
Yet, GrapheneOS also highlights a significant drawback: SafetyCore isn’t open-source and isn’t part of the publicly available Android Open Source Project. Developers express a lack of trust because open-source systems usually allow better scrutiny and transparency.
Google claims that SafetyCore provides a private way to classify content on devices and that users have control over this feature. It only scans specific content when a user enables the feature through an app. However, many users are doubtful about these assurances, especially since they weren’t informed of SafetyCore’s presence in the first place.
As noted by ZDNet, the main issue is that Google didn’t inform users about this service being added to their devices. If someone has a new Android device or has updated their device since October, they likely have SafetyCore installed without knowing. A major concern is that this service was added silently to Android devices without the explicit consent of the users. This has ignited worries around privacy and control over personal devices.
If you’re someone who doesn’t fully trust Google and is concerned about how SafetyCore operates, you can disable it. To do this, go to the ‘Apps’ settings on your phone, tap on ‘SafetyCore’ under ‘System Apps,’ and you should find options to uninstall or deactivate the service.
These recent events provide important lessons for both Apple and Google. If they want to enhance our phones with AI technology, they need to be upfront about what they are doing before it happens. Users deserve the opportunity to give their consent, rather than being kept in the dark, as this leads to unnecessary fear and skepticism regarding their privacy.
