[pyuser583]

The child is mentally ill and being treated clinically. Medication, therapy, etc. I’d rather not go too into the clinical details, because I’m not an expert, I’m a parent.

Roughly they are having difficulty with emotional regulation and processing previous trauma.

They’ve attended treatment camps in the past - which were very positive exercises for them. The approach to the internet at these camps is almost exactly what I am proposing: white lists and immutable history. They have to brew their own software but it’s done. I don’t see why it can’t be offered more widely.

Speaking more broadly, what controls do you think parents should have over adolescent Internet viewing?

Keep in mind most of the really bad stuff is on mixed content domains . Nobody builds a “watch-monkeys-die.com” website. They build a watchMonkeysDie tumblr or subreddit.

My case is not the best example because of the legit mental health concerns. But the principals are the same. As a parent I have decided my child needs granular access to the internet - how does that happen?

[soulofmischief]

I can't help you with respect to the library, that needs to remain a safe space for those who need it. But at home, do you have control over their networks or devices? I agree that the hardest yet most important things to moderate would be mixed content or UGC domains. Outright blocking some of these domains of course does the child measurable social harm, and social isolation is something you do not want when dealing with a problem like that. And the problem with just sniffing the network is TLS is going to obscure what content is being accessed.

If existing on-device controls are not adequate for you, I think one obvious modern approach which compromises on privacy in a situation might involve having a privacy-oriented on-device small language model look at the contents of each website and only alert you to suspect URLs, letting the child retain privacy if they stay away from toxic content. While this approach has many advantages, it also has disadvantages, such as requiring integration on the device level since an add-on wouldn't prevent a child from using another browser. But then we're moving into the territory of platform-locked devices with hardware attestation, age-gating etc. and the problem gets incredibly difficult fast.

Moderation on the internet is such a hard problem from many angles, and very unsolved. Multimodal models probably are the only way out; even now, the only reason platforms like YouTube, Facebook, etc. are usable by the general public is because tens of thousands of relatively poor individuals who are subjected to the worst material imaginable for 40 hours a week. It's a classist system as currently designed. A lot of these problems are very nuanced and need a lot of careful thought.

I do worry that if enough programmers who care about privacy and free access to information don't organize and provide stronger solutions, regulation will come from outsiders and it will suck. But right now, the current efforts to age gate the internet are concerted and not a rational measure against a lack of internal regulation; rather they're just exploiting that narrative in order to push through this legislation with broader public support.