Sure but things could be done to ameliorate the situation.
This is just spitballing but you could have opt-in and opt-out lists whereby if a certain threshold opt something out then it becomes opt-out by default thereby requiring some type of agency to be exercised for more controversial material.
You could also store things using some distributed encryption whereby multiple parties have to be online for the content to reassemble in the clear and only on the viewers machine.
You could add append style annotations so lies and misinformation at least has the opportunity of being challenged
You could have a more general region specific illegal content opt in for each node that has to be explicitly reaffirmed by the user, say every 30 days.
I'm not saying ban illegal content in the same way that the existence of stabbings doesn't mean you should ban knives, just make sure innocent people don't get flagged for it.
> you could have opt-in and opt-out lists whereby if a certain threshold opt something out then it becomes opt-out by default thereby requiring some type of agency to be exercised for more controversial material.
This would be an immediate DoS/censorship mechanism. Trolls would create a bunch of nodes and then have them all opt out of something they want to DoS so it falls out of the network.
> You could also store things using some distributed encryption whereby multiple parties have to be online for the content to reassemble in the clear and only on the viewers machine.
This is already how some of these systems work. The data distributed on the network is encrypted and to download it you get an identifier (e.g. a content hash) to locate it with along with the decryption key. The decryption key is only a few bytes and it can be included in the equivalent of a hyperlink. Without it anyone hosting the data can't read it.
Mega does something similar to this by putting the decryption key in the URL fragment so it isn't sent to the server but then client-side javascript has the key to decrypt the content with. This has poor security properties in their specific implementation because the server could be serving malicious javascript to extract the key, but new custom protocols don't have to allow that. Moreover, it might be worth something in terms of keeping attackers or insiders from snooping on their customers' data because someone with access to stored content might not necessarily have access to inject malicious scripts into client pages.
> You could add append style annotations so lies and misinformation at least has the opportunity of being challenged
This operates at a different layer, e.g. you could have a browser that does this and people could use it if they want to without building it into HTTP.
> You could have a more general region specific illegal content opt in for each node that has to be explicitly reaffirmed by the user, say every 30 days.
Then you would need someone to maintain all the lists even though it would be an exercise in futility rather than resulting in effective censorship of the material, since it would only cause it to be hosted in some other jurisdiction which every client could still transparently access as if nothing happened.
The problem here is that you either have an effective censorship apparatus or you don't. As soon as you have one, Saudi Arabia wants to enforce its blasphemy laws and China doesn't want anyone talking about Tank Man, which means you don't actually want one. Building half of one is just assisting the villains who want to build the whole thing.
> I'm not saying ban illegal content in the same way that the existence of stabbings doesn't mean you should ban knives, just make sure innocent people don't get flagged for it.
Prosecutors generally know the difference between "this is a crime ring" and "this is a Tor exit node". However, some of them are schmucks. The only real way to fix this is to make sure laws and courts don't allow them to ruin innocent lives, because the details of the technology aren't going to matter when someone who shouldn't be in a position of power has a vendetta that goes unchecked.
None of this is censorship. If people want to look at dirty photos then I honestly don't care, it's none of my business. I would, however, like to be able to run Freenet without those being served from My computer.
This is just spitballing but you could have opt-in and opt-out lists whereby if a certain threshold opt something out then it becomes opt-out by default thereby requiring some type of agency to be exercised for more controversial material.
You could also store things using some distributed encryption whereby multiple parties have to be online for the content to reassemble in the clear and only on the viewers machine.
You could add append style annotations so lies and misinformation at least has the opportunity of being challenged
You could have a more general region specific illegal content opt in for each node that has to be explicitly reaffirmed by the user, say every 30 days.
I'm not saying ban illegal content in the same way that the existence of stabbings doesn't mean you should ban knives, just make sure innocent people don't get flagged for it.