Verify your AI. Microsoft is set to add branded footers and mandatory watermarking to Copilot to stop phishing and secure sensitive data.
Microsoft is adding two new administrative controls for organizations that use Microsoft 365 Copilot with a focus on branding and transparency through watermarking. With these updates, it will allow enterprises to verify the authenticity of its internal Copilot environment and label AI-generated multimedia content.
One of the changes that Microsoft will implement is the ability for admins to insert a corporate logo into the footer of the Microsoft 365 Copilot app beginning in late March. This option is being introduced to help reassure employees that they are using an official, secure, work-managed instance of Copilot where they can more safely input sensitive data.
Microsoft has also updated the timeline for a mandatory policy setting that applies visual and audio watermarks to AI-altered. This transparency feature was originally set to launch in February, but it’s now planned for arrival in the Cloud Policy service by mid-March.
The branding update will use the existing organization theme configuration within the M365 Admin Center to load the required logo. Right now, only the logo asset will be applied to the Copilot footer, but other theme elements, like custom colors will not be available.
The watermarking for video and audio content, meanwhile, will need to be manually enabled by administrators through the “Include a watermark when content from Microsoft 365 is generated or altered by AI” policy. Once this is activated, markers will be applied to assets such as videos created in Clipchamp or audio overviews generated from Word documents.
Static image watermarking is handled through a different mechanism that became available to users in late February. By separating them, Microsoft has made video and audio watermarking governed by tenant-wide policies, and image watermarking a user-level preference, at least for the time being.
With these improvements, organizations can greatly reduce the risk of phishing by providing a branded Copilot interface. The introduction of watermarking also helps businesses meet emerging compliance and ethical standards surrounding synthetic media – this can help to prevent the spread of deceptive media.
