Hacker News new | ask | show | jobs
by brucethemoose2 1060 days ago
I hope it doesn't get killed by CSAM spam or stuff like that from trolls.

Automated federated filtering is not impossible. In fact, an distributed setup (where volunteers host image/text classifiers) like AI Horde already does seems pretty doable.

2 comments

Unlikely. The BBC would be very daft to run an open-signup server. And a closed-signup server would only have the problems that BBC employees themselves bring.
But won't replies from other servers show up on the BBC's page if they post something?
Not necessarily, no.

There's a whole long standing organization of sysops dedicated to silencing and blocking the (few, I've never even heard of one myself) places on the WWW that the sort of stuff that you are talking about would come from. If the BBC sysops are on the ball, they'll already be looking at what block lists to subscribe to.

In fact, they've had the system up since June, and I wouldn't be surprised that they haven't already been importing blocklists. Alas, the "about" page doesn't say, as it normally would, what servers are moderated. But I can understand why the BBC would not want to get into the silly game that would (and does) result if it did publish its moderation list.

It's not well publicized, and resources for new sysops being poor is one of the problems with Mastodon et al., but I hope that the BBC people are on the ball in this respect.

* https://writer.oliphant.social/oliphant/the-oliphant-social-...

> There's a whole long standing organization of sysops dedicated to silencing and blocking the (few, I've never even heard of one myself) places on the WWW that the sort of stuff that you are talking about would come from.

I was thinking more along the lines of subreddit trolls, who will come in and spam shock/csam images trying to bait the mods, not actual CSAM distributors with known IPs or whatever.

The BBC will certainly attract that kind of internet denizen. They would have to have employees (or automated system) staring at their post replies constantly, watching for this kind of stuff.

The FediVerse doesn't work that way. There is no "the" mods.

The people posting this stuff won't be posting to the BBC. They will be posting to far away places each with their own moderators. They'll either get suppressed at the source, by those moderators, as complaints come in (which they will probably do quite rapidly); or it will get around that those far away places don't moderate, and the far away places will get silenced and blocked and make it onto the aforementioned lists, meaning that the (civilized) world won't pull any posts from them.

The reporting system, bear in mind, reports to the origin server, which won't be the BBC. If I, for example, see an unacceptable post in reply to something posted by the BBC, my report doesn't get seen by anyone at the BBC. It gets seen by my local sysop and the sysop of the origin node.

Hm. Interesting, thanks for elaborating.

If spamming smurf accounts on the reputable instances is hard enough, I guess this isn't an issue.

I clicked on their federated timeline and the first thing I saw was a post from someone named "horsefucker thicc".
TFA says:

> Unlike most Mastodon servers where you can sign up for a personal account, we're only using this instance to host BBC accounts; it’s a place for us to publish in the Fediverse. If you have a Mastodon (or other ActivityPub) account from another server, then you can easily follow our accounts.

So if there's CSAM, it's coming from the BBC itself, which is hopefully unlikely.

And the fediverse / Mastodon already takes care of filtering. Niche instances failing to police CSAM are defederated by mainstream instances, which means they're unreachable to anyone on the mainstream instances. This could be improved and automation tools for use by moderators are certainly welcome, but generally speaking as a user you're not going to see objectionable content unless you go looking for it or your admin is negligent.

edit: typo