The Content Moderation console displays detection results from the Content Moderation API. You can manually review these automated results to improve their accuracy for your business. After you complete a review, the service applies your decision to all future detections of identical content. This topic describes how to use the Audit feature.
Background
By default, the Audit page shows only content that the system flagged with a
revieworblockstatus. To include content with apassstatus, you can adjust the settings in the console. For more information, see Customize policies for machine-assisted moderation.You can manually review images, videos, audio, and text. However, the service automatically adds your decisions for images and text to the corresponding sample library.
Detection data is retained for a maximum of seven days. Review your content promptly to avoid data loss.
Procedure
Log on to the Content Moderation console.
In the left-side navigation pane, choose .
Click the tab for the content type you want to review: Image, Video, Text, or Audio.
Mark the content as follows.
NoteThe following steps use marking images as an example. The process is similar for other content types.
For acceptable content that was flagged with a
blockorreviewstatus, mark it as Pass.For violating content that was flagged with a
revieworpassstatus, mark it as Block and select a reason: Porn, Terrorism, Ad, or Undesirable Scene.To process multiple items at once, select them and click Batch Mark Pass or Batch Mark Block.
Click an item to view its Details.
The service immediately applies your decision to the reviewed item and any identical content. It also automatically adds the item to the appropriate sample library.
As shown in the following figure, the red System label indicates a system-generated library for reviewed items. Text libraries function similarly.