Microsoft Purview | Information Protection: Feedback Loop for Content explorer and DLP alerts (GA)
Microsoft is introducing a feedback mechanism in the compliance portal to enable administrators to report false positive results for sensitive information types (SITs) and trainable classifiers found in Content Explorer and DLP Alerts (non-legacy alerts). This Feedback Loop is designed to enhance the accuracy of SITs and classifiers and will allow Microsoft to make systematic improvements to reduce false positives. The feature is currently available in public preview, with rollout expected to be complete by the end of December. By offering a compliant feedback channel, the new Feedback Loop enables administrators to share their feedback on policy mismatches seamlessly, avoiding the need to submit examples offline. The new feature is supported for newly created or updated items and will soon be generally available for Content Explorer and DLP Alerts. In the compliance portal, administrators can easily review and mark classifier matches as true or false positives, and there is also an option to submit redacted copies of falsely matched items to Microsoft with feedback to improve classifier accuracy. On the SIT overview page for a given classifier, administrators can also provide feedback and see a summary of feedback submitted on the Feedback results tab. This feature will be beneficial for enhancing the overall accuracy of SITs and classifiers and enable Microsoft to make systematic improvements to prevent false positives in the product.
The post Microsoft Purview | Information Protection: Feedback Loop for Content explorer and DLP alerts (GA) appeared first on M365 Admin.
Published on:
Learn moreRelated posts
Microsoft Purview compliance portal: Information Protection – Trainable Classifiers in Server Side Auto-labeling
If you work with exchange, OneDrive, or SharePoint, you'll be pleased to know that Trainable Classifiers are now available for server-side aut...
Microsoft Purview compliance portal: Endpoint Data Loss Prevention – Contextual evidence match for endpoint DLP
This post talks about a new update on the Microsoft Purview Compliance Portal that allows you to see the context and sensitive information typ...
Microsoft Purview compliance portal: Communication Compliance Advanced trainable classifiers to detect workplace safety
Microsoft Purview Communication Compliance is set to introduce advanced trainable classifiers that leverage large language models to detect wo...
Microsoft Purview compliance portal: Information Protection – Keyword highlight support for document trainable classifiers
Microsoft Purview's latest portal update includes a highly requested feature that highlights keywords and phrases that contribute significantl...
Microsoft Purview compliance portal: Insider Risk Management – Microsoft Information Protection (MIP) ML Classifier support
The Microsoft Purview compliance portal now has Insider Risk Management support for Microsoft Information Protection (MIP) ML Classifier. This...
Microsoft Purview compliance portal: Information Protection – SIT Feedback Loop for Content explorer and DLP alerts (U.S. Government clouds)
Microsoft's Purview compliance portal now allows admins to provide feedback for potential False Positives for sensitive information types (SIT...
Microsoft Purview compliance portal: Information Protection – Improved Resume OOB Built-in Trainable Classifier
Microsoft Purview compliance portal's Information Protection just got an upgrade; the Resume Classifier is now equipped with better technology...
Microsoft Purview compliance portal: Information Protection – Contextual summary support in service-side auto-labeling
Microsoft Purview compliance portal has introduced support for contextual summary in service-side auto-labeling. This feature enables users to...
Microsoft Purview compliance portal: New and enhanced Source Code trainable classifier
If your organization relies on the Microsoft Purview compliance portal to manage data loss prevention policies, be on the lookout for the enha...