Comment by vaylian
21 hours ago
You are correct in that both option 1 and 2 are possible. For end-to-end encrypted messages only option 2 is possible. The content will be scanned directly on your own device and the data will be sent to the authorities without your knowledge, if the software detects something suspicious. This is called client-side-scanning.
The proposed Apple system was at least more restrained in that it was looking to identify known abuse images. Which is better than the Google one which aims to identify new unseen content which constantly flags parents acting legally sharing photos to medical professionals.