[AnthonyMouse]

> Its a second order effect. The problem is predators get children to take pictures and distribute them.

It isn't. The solution is obvious. You punish the predators and not the children.

> My understanding of what is being proposed is that cameras will have basic nudity detection on them, and they will refuse to take the pictures if the device is registered to a person under the age of 18.

Which seems both problematic (now every device needs to be registered?) and ineffective (the predator has the children use a device registered to an adult or an older device without any such constraints).

> The central point that you are missing here is that something which was quite hard to happen in 1999 was destructive to a child life. Now its much easier to do, and there is an actual economy in sourcing, exploiting and trading these pictures/videos.

This seems to be your central premise but it's also not even true. If a predator got children to take such pictures with film camera in the 20th century and was then in possession of the negatives, there was no centralized system to detect this or prevent them from having a darkroom to make and distribute copies. The solution is to have the police arrest them, which continues to be the solution even now without needing to compromise the devices of every innocent member of the public.

> But if we want to have strong controls over your data, then sharing should be controlled by the owner, not the platform.

You're assuming the conclusion -- that there should be a platform in a position to control (i.e restrict) sharing. If communications are end to end encrypted, no one other the parties even knows what it is. That is what you are by implication prohibiting -- unintermediated private communications.

[KaiserPro]

I am deeply worried about privacy. the problem that I have here is if we don't provide a sensible middle ground, we will get a terrible solution imposed (ie the age gate spyware, instead of the rollingo ut of the mobile block lists the UK already had)

> It isn't. The solution is obvious. You punish the predators and not the children.

the core problem is that in the UK more and more CSAM is being prodiced by kids themselves. Most grooming of digital images is done remotely, via exploitation. (ie I have pictures of you, please send more or I'll send them to x)

Now, because the person extorting them is not taking the pictures they had a defence of "they were sent to me, I didn't know what to do so deleted them blah blah blah"

These people hunt down the kids sending stuff to lovers and the like, that gets leaked, because kids are cruel, and try and extort them.

I would gently ask you to look in the changes into sentencing guidelines and the evidence used to compile that advice.

Yes, older devices can be used, and oncein the hands of extortionists then they are in deep shit. The point of this, and the point that Jess Phillipson was getting at, is if kids can't take these pictures as easily, they can't be leaked as easily which means there is much less (but not zero) chance of being exploited,

> You're assuming the conclusion -- that there should be a platform in a position to control

Sorry my bad phrasing.

the ideal solution here is end to end encrypted but also encrypted image but with the keys controlled by the owner of the image "private" computing is the answer, where every view of private images requires a key exchange. there are hardware locks to stop people getting at the frame buffer, only apps that you approve can have access, and flagged data that you send auto deletes and expires.

Ironically this is basically extreme copyright enforcement(everything you create, more or less is your own copyright)

But we can't get to that yet, because Ad tech/meta/google doesn't like that.

TLDR its less obvious but I also share your worry about privacy.

[AnthonyMouse]

> the problem that I have here is if we don't provide a sensible middle ground, we will get a terrible solution imposed (ie the age gate spyware, instead of the rollingo ut of the mobile block lists the UK already had)

The problem here that if every year you're asked to make a new compromise between where you are and a fascist police state, you're only traveling in a single direction on the road to hell.

To actually stop them you need to turn around and go the other way. Dedicate your efforts to removing the existing authoritarian surveillance apparatus so that they have to spend their resources to stand still instead of you, so that failure is "it stays the same" and success is "authoritarianism declines" instead of the unacceptable premise that the only options are making it worse or making it worse even faster.

> Now, because the person extorting them is not taking the pictures they had a defence of "they were sent to me, I didn't know what to do so deleted them blah blah blah"

Any kind of sensible court is going to both take evidence of the extortion as an invalidation of that defense and also convict them of the extortion.

Moreover, "it's sometimes hard to punish the perpetrators" is a ridiculous rationale for punishing the victims in their place.

> But we can't get to that yet, because Ad tech/meta/google doesn't like that.

We can't get to that because that doesn't actually work, for multiple reasons in this context:

The first is the same reason every piece of Hollywood content is in the hands of pirates despite Disney et al demanding useless DRM. It simply doesn't work; a single vulnerability allows every piece of content to be extracted and once extracted by any one person it can then be distributed to an unlimited number of others forever. The entire premise is snake oil. The reason Hollywood wanted it to begin with, and Google and Apple adopted it to use against them, is the same -- it's hopeless at preventing copying but it enables the monopolization of devices because then unapproved competing devices are illegal.

The second reason is that the person in an image is not necessarily the person who owns the device, and the device has no way to know that. It does you no good for the "owner" of the image to control it when the device thinks the "owner" is the predator.

And the third is the pragmatic issue. People expect to be able to take a picture and share it with whoever they want. You take a picture of yourself and text it to your mom, both you and she expect her to be able to email it to friends and family etc. You take a picture of a defective mannequin to file for RMA and you expect the customer service rep to be able to upload it into their internal RMA system or forward it on to the manufacturer in another country. If the default doesn't allow this then people will justifiably object. But if it does then 99% of people will never change the defaults or even realize that they can, and then the proposal is worthless because the default everyone wants for the common case is exactly the thing you don't want in the problem case, since then the default allows the picture to fall into the hands of a thousand strangers and devices that aren't in the authoritarian ecosystem and you can't get the cat back in the bag.