Facebook is asking to use Meta AI on photos you haven’t yet shared | Hacker News Reader

Facebook is asking to use Meta AI on photos you haven’t yet shared

techcrunch.com

91 pointsabsqueued16 comments

[dupe] More discussion: https://news.ycombinator.com/item?id=44401406

Comments moved thither. Thanks!

There's a big active thread about this here already: https://news.ycombinator.com/item?id=44401406

I think the TechCrunch headline is slightly more accurate than the Verge headline, which is "Facebook is starting to feed its AI with private, unpublished photos".

In both cases they imply training models is happening when that's not been confirmed.

(Facebook could help here by answering press inquiries about it, which they apparently have not done.)

While I agree the articles are click bait since this as not been confirmed, it's not far fetched to assume that a giant corporation with a terrible track record and a big legal department worded their TOS like this because they intend to use that capability.

Exactly.

Given Facebook is about as voracious an actor of surveillance as has ever existed, their track record of respecting few red lines until they have been caught crossing them egregiously, and the bright spotlight Zuckerberg is currently shining on their AI ambitions, it defies reality to imagine them forgoing any data they can get there hands on.

You just know there is a dashboard that summarizes all potential data sources, and engineers wake up with the shakes and sweats, after dreaming that Zuck was standing behind them, with furrowed brow, and pointing to a stat that shows 2% of someone’s most private information still hasn’t been plundered.

Ok, a little hyperbolic. But he & Meta are relentless.

Yeah, I read the title of the other one and thought it was just about pictures you sent through Messenger or put in private on Facebook, so I wasn't too bothered (because I assumed they'd do this already), but actively reading your camera roll is next level.

> but actively reading your camera roll is next level.

Now that it was established that they wrote malware to bypass tracking protections, nothing surprises me. Apps written by Meta are malware, as far as I'm concerned.

> In both cases they imply training models is happening when that's not been confirmed.

It is safest to assume that your photos are being used for training.

I like how the headline is truncated. I was thinking of 'seen' or 'taken' but turns out it ended with 'shared'.

What do you like about it?

How is it the photos are shared when they’re not shared?

Perhaps you're not aware, but starting a sentence about something you don't like with "I like how..." is basically a form of sarcasm.

That's true, and people should learn to recognize it. But in general, sarcasm is easily misunderstood in pure text. You read it with a tone in your head, but they can't hear it.

Also, it's best to avoid it a site like this with many non-native English speakers. It's an extra layer of difficulty.

Ok, good info. So GP means to evoke the type of information shearing garbage some of us are wise enough to expect from unaligned and underspecified human-emulating but self-serving autonomous digital systems, and not comment on their sincere affection for information loss in clickbait titles that. . .

In this case it might not have been sarcasm. The headline cut in just the right place to let you fill in the blank for your own amusement.

Soon with their AI, it'll be "photos you haven't yet taken"

Could be more fun options:

  - Photos you haven’t yet known you’re in. 
  - Photos you haven’t yet deleted. 
  - Photos you haven’t yet thought better about taking.

Reply on news.ycombinator.com