Hacker News Reader

Comment by briefcomment | Hacker News Reader

I may consider getting a stand alone camera now.

Apple is going to scan images you download too, not just ones you take with your camera. Leaving their ecosystem is way more important than just getting a standalone camera.

Do you mean this would apply to Macs too? So if I used some non-Apple photo app, and uploaded photos from a camera, they would be similarly monitored?

No, they wouldn't. It's only scanning photos you're uploading to iCloud Photo Library via Apple's Photos app.

> Do you mean this would apply to Macs too?

From [0]

'These features are coming later this year in updates to iOS 15, iPadOS 15, watchOS 8, and macOS Monterey.'

So yes.

Also by the OP:

> iMessage encrypts messages end-to-end by default; however, if you have iCloud backup turned on, your messages can be accessed by Apple (who has the keys for iCloud backups) and, by extension, law enforcement with a warrant.

So if you use iMessage and iCloud Photos then they will be scanned by Apple. Won't be surprised to see the on-device scanning feature more prevalent on M1/M1X Macs.

[0] https://www.apple.com/child-safety/

Ah thank you.

So if you use iMessage and iCloud Photos then they will be scanned by Apple--> it is a lie. This is definitely not the case.

Yes, with iCloud backup turned on, messages can be accessed by Apple, that hasn't changed, has been the case for a decade. It is specifically written in the technical document that they are not scanning the iMessages photos in the cloud the same way that they would scan your iCloud photos.

One thing that might happen is that they could MITM your internet connection by default while maskerading it as a privacy feature.

https://www.theverge.com/22573519/apple-private-relay-icloud...

Even if it did, iCloud is less of a necessity with a Mac than it is with an iOS device. I have a little more agency and utility on Macs.

Oh... you just reminded me of the upcoming VPN service that will act as a proxy of sorts for internet browsing. If they do MITM HTTPS then they will basically scan evertything about you and your internet habits.

When did it became like that? Their only saving grace was that they were okay-ish with user privacy.

If we are going to be totally paranoid then a lack of pictures on your mobile phone could be seen as an indication that you have something to hide!

Sure, then let them play their authoritarian hand in attempting to follow through on that. Maybe at some point, it gets ridiculous enough that some critical mass of people get fed up and we are able to reverse course.

Nothing to hide right? As an otaku who collect figurines I can't wait to have the police knocking at my door because of a false positive triggered by a 1/8 scale statue of Megumin casting a fireball.

Cameras in the future are not immune to what we may see on upcoming smartphones. For example, Sony revealed an image processing sensor with compute capabilities built-in. It's able to not only send photon measurements but send metadata on what's being captured in the image. https://techcrunch.com/2020/05/14/sony-shows-off-first-combi...

Remember when scanners would allow you to scan banknotes, even those with a Eurion code?

I suspect we'll be seeing more of this kind of thing in digital stand-alone cameras. GPS-enabled? You just lost the ability to photograph demonstrations. Trademarked logos i the background? Maybe the shutter won't click.

Scanners with banknotes remind me of my kid's LeapFrog book. It's basically a book and a pen which can read whatever you tap against. It has an optical sensor at the tip of the pen and the print just has a bunch of microscopic dots barely visible to the naked eye. https://en.wikipedia.org/wiki/LeapFrog_Tag

The way it works is people report CP to a third party organization that fights CP. A hash of some sort is recorded in a central repository that is used by many services including every major social media company. Those companies then use it to delete/ban/etc users who post images matching those hashes. Generally, they also work with law enforcement.

In the current iteration they are not using AI or anything to identify "bad" content. They are matching hashes for files on on your device against known bad hashes/content.

I'm not trying to convince you one way or another about it, just clarifying what is happening.

> In the current iteration they are not using AI or anything to identify "bad" content. They are matching hashes for files on on your device against known bad hashes/content.

This is misleading as these are not what most technical people would think of as a hash. These are "perceptual hashes" that can compute a distance, instead of a boolean equal/different, leaving space for false-positives instead of extremely rare hash collisions.

https://www.hackerfactor.com/blog/index.php?/archives/929-On... https://rentafounder.com/the-problem-with-perceptual-hashes/

Reply on news.ycombinator.com