Microsoft 365 begins watermarking AI-generated content

Microsoft is introducing a new policy in Microsoft 365 that allows organizations to watermark content created or edited using AI tools such as Copilot, Windows Central reports.

Initially, the policy applies to audio content created or edited using AI in Microsoft 365. Video support is expected to launch in March 2026. The feature is not enabled by default and must be manually enabled by administrators. The watermark cannot be customized: the text and location are predetermined by Microsoft.

Even if an organization chooses not to use visible or audible watermarks, AI content will still be marked with metadata. This information can show which AI model was used, which application created the content and when it was generated, among other details. AI-generated images are processed separately through a different user-level policy.