Over the past three years, I, along with a team of other people have created an application called SmokeDetector. Some of you may have heard of it. For those who have not, here's a 30-second run-down.
SmokeDetector scans all new and updated content, network-wide, trying to find spam. Over time, it's become very accurate at doing this, and has a team of people continuously improving it.
At the start of this year, we began setting up a project to make use of SmokeDetector's accuracy to flag spam automatically. Full details of how this works are available in this post on Meta.SE, but again, a brief version: we select combinations of patterns that have a history of being accurate, and check them against a database to make sure they're at least 99.5% accurate. Once those combinations are satisfied, we cast up to 3 spam flags on the post automatically.
We believe we've designed this system to be both accurate and safe, but now we'd like your input on it. If you have any comments, questions, concerns, or objections to this system, please post them as an answer either here, or on the Meta.SE post, and we'll do our best to answer them.
Note