While reviewing suggested edits and first posts by new users today, the review system did something new - an alert appeared that suggested that what I was reviewing was probably spam, and prompted me to take a careful look at it.
Should I be doing anything differently than I otherwise do while reviewing when I see this? Additionally, are my actions contributing to helping the spam prevention systems learn?
We recently strengthened spam protection in the area of suggested edits. As multiple spam networks tend to target the same seemingly innocuous posts, we've started keeping track of how many spam edits any given post receives over a period of time, and now deal a severely harsh penalty to those that attempt to spam known targets.
Nothing identifies these from non-targets, a spammer has no way of knowing what's being watched more closely.
As this tends to come in in floods of up to hundreds of edits at times, a bit of cleaning up is needed after the spammer has been blocked. Generally, most are blocked after only a couple of attempts have been reviewed, but what remains is everything else that got in prior to the block being enforced.
Think of it a bit like email, where you see a sudden surge of spam, then your filters learn and stop more from coming in. You've still got some in your inbox that you have to throw out, and that's what you're doing when you see this.
If you see this message, the origin has been, or will very shortly be blocked (99.9% of the time, they have been). The system is just alerting you to the fact that you're probably looking at spam, so make sure to review it extra carefully.
So, what do I do with it?
If it's spam, flag or reject it as usual - no other action is needed.
If you are certain that it is not spam, and possibly just an edit that was caught in the metaphorical cross fire from a user that might simply have an Internet provider in common with the spammer - then simply accept the edit if it's good, or reject it for reasons other than being spam.
It's always (or almost always) accurate, why not just delete it when you block them?
We'd like to get to that point eventually. First, there are more opportunities to keep more of it from ever entering the system, and we want to see what those do as far as what actually gets in front of reviewers after a large coordinated attack, such as what we've been seeing for the past ~10 days.
Then, while our error rate so far has been extremely low (less than 1%), it is a concern. The idea of this system is to interfere with sincere use of the sites in the most minimal way possible.
I'm helping the system learn?
Yes. By rejecting or flagging, you're sending the system additional signal. If they're on the verge of one of the various cascading blocks that we impose, you could very well bump them with your review.
By saving something that isn't spam, you not only salvage a good contribution, you help us to better analyze and improve the system overall.
Are these (posts/edits) otherwise altered by the system?
Nope, you're seeing them in their pure, gelatin coated state as submitted. The system doesn't remove links or images - this would just obfuscate what is spam to varying degrees and in the case of the rare false positive it would just needlessly mangle the contribution.
So, Pooh Bear, now what?
We have more work to do, I'm not going to be happy until we keep this crap mostly out of people's global in-boxes. However, this is a pretty good leap forward, and there's even more places we can catch it before it even hits the queue. Still, once that's done, you'll still see these on occasion.