I think (but am not sure) that there are long established postal laws in most territories about sending “obscene” material through the mail. I think this was used to prosecute pornography publishers in earlier times. BUT you needed to (a) intercept mail and (b) have a good reason and (c) get a warrant to open (interfere with) that mail.
Possessing pornography was a separate issue which may or may not be allowed. Typically (I think) authorities went after publishers not consumers - because they were easier targets to pin down.
Which would seem to imply that if you’re sending encrypted traffic at the request of a recipient the as a publisher of “obscene” material then unless you are delivering very clearly illegal content to a user then you should not prosecuted.
I haven’t got a single source for anything I’m saying, so I might be entirely wrong - I’m simply going off half-remembered barely-facts. So please do argue with me!
It's different, because you are willingly sending a reply to a known UK address.
In the website scenario, there are no physical addresses with a geographic component to them. The physical topology of the network is only known by the operators of the network. Only they know where the routers are physically located.
This means geoip blocking can only ever be done on a best effort basis. Actual blocking can only be done by the operators of the routers, which is why it is unreasonable to expect the website operator to be responsible for perfect compliance.
