Forum premoderation

I've bitten the bullet and decided to try pre-moderation on the forums.

From now on for any new forum contributor, there is a step where their entry ends up in a moderation queue. It won't show in the forums at all until it's been moderated and OK'd.

I hope it doesn't create too much work for moderators but I think considering it's just for new members, it's unlikely to be an issue. And the regular spam reports should drop to near 0 with this in place.

I will also be getting notified whenever something is added to the queue, so hopefully genuine posters' contributions are approved pretty quickly.

Sounds like a worthwhile move.

I'm assuming new members will get an explanation after (before?) posting their thread, that it might take up to a few hours before it'll show up?

I fear most of the fake support phone number scam spam will slowly shift to the blogs now, but maybe the subdomains will mean that spam is less valuable, so won't be quite as massive.

I still think custom protection on the signup form - varied each time they break through en masse, to waste extra developer time at the source - has the best bet of being a long-term solution.

Yes, they are shown a message that it may take a while to be reviewed. And they are sent another email when their thread has been accepted.

Shifting to the blogs is a concern. I don't really want to pre-moderate those and them being on subdomains does help isolate the effect of the spam especially because any new blogs are explicitly excluded from Google which is their main game. Might have to think of other options there.

The managed challenge on signup that I currently have really does a HUGE amount from what I'm seeing in the logs. With this last batch yesterday it really was either just humans clicking through (maybe assisted with a chrome extension or similar) or AI agents that any amount of custom protection probably would just as easily be bypassed. Pre-moderation future proofs against all of that and as long as it's not an overwhelming amount of work will also just avoid the spam appearing for even a short amount of time.

Just got a first questionable case in the premoderation queue. It's good old fashioned human astroturfing with a small percentage of being okay; I'd snip the company name and approve - but that's not an easy option in the current flow, so request to add that? Also a link to the profile of the account in question would be handy, in case of images which need to be deleted and a general overview of what else the account has done so far. Thanks!

[ Edit: Edited on 25 Jun 2026, 14:28 GMT by Sander ]

Another feature request for this: show the link destination in brackets on the premoderation screen (as on the forum for moderators) so we don't overlook sneakily hidden ones.

Yep fair point. I'll do a round of updates to this soon.

I'm struggling a bit what to do with AI-generated but otherwise non-promotional posts in pre-moderation. The purpose of pre-moderation is just stopping the worst spambots, for which it works perfectly. But of course I can't help but notice other things as well, and with AI-generated posts in particular there's always a small chance of being wrong.

Do I delete and contact the poster to let them know? That takes on a lot of overhead, more than I want when the forum in question is not one I'm responsible for.
Do I delete without further action? That removes the overhead, but sucks for the poster if I got it wrong, and will probably leave a lasting bad impression.
Do I let it linger in pre-moderation until the responsible moderator sees it? That makes the time in premoderation much longer than intended for this system.
Do I approve it as not obvious spam? But then if I saw the post on its own, I'd flag it for being AI...

Yeah, I see your point. I guess if it's enough of a doubt we have to leave it and treat it the same way we'd treat a regular post. If we're referring to the one that was in the queue, I decided to leave it for now. It's definitely suspicious I agree but there's not much to go on here other than the fact that it looks to be clearly AI.

We could add an option to delete it as "AI generated content" and send an automated message to the user notifying them of that?

The one you just approved was indeed the one which triggered this. I've deleted several others which were more clearcut, and just took the hit of doing notifications, but as this one had a marginal chance of me being wrong and thus triggering a back-and-forth, I was starting to wonder what I was setting myself up for.

Quoting Peter

We could add an option to delete it as "AI generated content" and send an automated message to the user notifying them of that?

I like that, yeah! Automate the overhead. Still with a human in the loop to make the determination, but with a better process afterward.

I mean, it should probably do that whenever the "Delete" button is hit anyway, as right now it pretty much will just leave them thinking they are stuck in the queue forever.