What they're not telling you: submitted by

Marcus Webb
The Take
Marcus Webb · Surveillance & Tech Privacy

# THE TAKE: privacy-assurances-after-stuffing-devices-with-their-ai-model.html" title="Google removes privacy assurances after stuffing devices with their AI model" style="color:#1a1a1a;text-decoration:underline;text-decoration-style:dotted;font-weight:500;">Google Photos' "Privacy Alarm" Is Propaganda for the Technically Illiterate Google announced client-side scanning. The tech media responded with performative panic. Classic misdirection. Here's what actually happened: Google added on-device ML processing before upload—standard hash-matching against known CSAM databases. This *reduces* server-side surveillance compared to their previous infrastructure. The real story? It validates what NSA contractors have known for years: you can't build consumer surveillance at scale without appearing to build consumer *privacy*. The optics matter more than the mechanics. What should alarm you: not that Google scans photos, but that we're debating *where* the scanning happens—device or server—as if location determines morality. It doesn't. Google processes your visual data either way. The announcement's genius wasn't technical innovation; it was reframing acceptable intrusion as protective infrastructure. The mainstream media missed the actual play.