Another angle here is that we can probably gather quite a lot of evidence about anyone who publishes child abuse material on nostr. Making it clear that we have this capability could act as a deterrent.

In Nostr, I've only seen CP created by neural networks. As long as they don't use materials with real children, I don't care. On the other hand, in a number of countries, even a picture of a naked ninja turtle will be taken by the court as child pornography. And relay admins are really risking their freedom. That's a problem.

Yeah, I think many countries would consider ML generated CSAM to be the same thing as an actual picture or video taken of a sexualized child.

UK for example: https://www.bbc.com/news/uk-65932372

And Australia it can be text, not just images or video: https://www.lexology.com/library/detail.aspx?g=be791d54-9165-4233-b55a-4b9dad5d178d

The risk for most relay operators is that people will using nostr for the discovery / connection between people who then go in to other apps / servers to actually exchange the content. Apparently it’s a kind of whackamole with different hashtags people search for this kind of content.

This is great. Now we need people to go do this stuff in a way that ideally any client and relay can use. Also places like Nostr build. One idea would be to have a way of indicating which relays are serving this stuff. Metrics like that could potentially incentivize big relays to get this done.

Someone very dear to me is about to join a research team working on this subject. I really rather wish she wasn't tbh.

Good read, certainly opens my eyes to things I didn't know were going on.

That said, I have a hard time grasping how any node operator could implement the recommendations suggested.

Big Tech spends millions to police this stuff, many of the tools needed are financially out of reach. Even PhotoDNA which is free, is only available to "qualified" organizations.

I fear that this gatekeeping of CSAM moderation tools will be used to justify the continued centralization of the internet.

"Think of the children" has long been used as an excuse to infringe on our rights. Big Tech won't hesitate to use it to squash #Nostr in it's infancy. Too many $$$ at stake.

I was just saying today that I predict that the disenfranchised "alt right" who have been frozen out of social media now for about 5 years, at least, will adopt Nostr fast and make "goggles" to filter their view.

The great thing is that while you can do this to Nostr, it's simpler and easier not to, and thanks to hidden services, and soon that will include Indra, there will still be unimpeded transit of messages that would otherwise get flagged as "nazi" or "alt right" or whatever, will still exist.

They are gonna still control a lot of things via DNS, and if you are paying attention, that means that pretty soon, give it 2 years, most of Nostr will be walled silos with tiny conduits between them, and a small, underground network of relays that don't presume to curate the traffic.

I just don't accept that liability should be on relays. Period. This is never going to be ok and it's why although there will be a brief flowering, the ways to lock down most of Nostr are already ready to apply.

分散型ソーシャルメディアプロトコルにおける児童性虐コンテンツの管理に関する論文らしい。

だいたいはFediverseへの言及で、Nostrについてはまた別の問題があると述べられているのみみたいだけど、リレー管理者やメディアのホスティングをやっている人たちは何らかの対処をする必要があるだろうと。

nostr:nevent1qqs28y0thztgptz7udkqluvgvs4ctneerxxg0uj7uq5lxqwsfugnm3spzamhxue69uhkv6tvw3jhytnwdaehgu3wwa5kuegzypmvwx4w8fy378v7a3ruhgt7y2wd5sgn5zamde4wzamdwep798905qcyqqqqqqgynaav2