You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Similar to my issue not being able to distinguish between my machine-created reports, being able to set a report threshold after which a post gets escalated would be supremely helpful
The text was updated successfully, but these errors were encountered:
I think we'd probably lean towards achieving this via an automod bot, which would look at the stream of events coming from the ozone backend, keep track of counts, and escalate subjects based on that. Our design philosophy on this is to generally keep the ozone "state machine" (of transitions between subject review states) simple and consistent, and use flexible bots to automate flows more. This should become particularly powerful when we start to lean in to the tagging system more.
As an example, we currently have appeals go directly to the escalation queue, but I don't love that and sort of wish we had that behavior driven by a bot, not by the ozone backend itself.
As a bit of context, the way we use "Escalation" is for hand-off between team members, or as a bucket for content which will require group discussion, not necessarily for more-urgent or higher-harm content.
I feel like what you are actually trying to get at with both this request and the #84 request is the ability to re-order the review queue by priority, and give human reports higher priority over automated reports. Eg, have them show up at the top of the queue, instead of just updated-at time sorting. Would that work for you? We'd like to get that in eventually ourselves, for high-harm or "trending-reported" content to "jump the queue".
Similar to my issue not being able to distinguish between my machine-created reports, being able to set a report threshold after which a post gets escalated would be supremely helpful
The text was updated successfully, but these errors were encountered: