Correction dated August 9, 2026: We could not find an official Microsoft announcement for watermarking AI-generated or AI-edited content specifically within Microsoft 365 Copilot. The earlier version of this article presented the capability as announced; there is currently no reliable primary source for that claim.
What is officially documented today
As of the date above, the public Microsoft 365 roadmap and Microsoft documentation do not contain a corresponding Copilot announcement. Other Microsoft capabilities using the term “watermarking” relate, for example, to protecting sensitive content and should not be confused with labeling AI-generated content.
What organisations should do
- Do not design compliance or approval processes around an unconfirmed Copilot capability.
- For now, govern disclosure of AI-assisted content through binding policies and templates.
- Check new roadmap items against the corresponding Microsoft documentation before relying on a technical control.
Official reference
We are happy to discuss how to make Copilot use in your organisation secure, transparent, and traceable as part of a Microsoft Security Assessment – get in touch with us.
Govern Copilot transparency reliably
Define disclosure and information-protection rules without relying on unconfirmed capabilities.
You speak directly with Sebastian Kerssen – working in the Microsoft ecosystem since 2014, with personal consulting and a dedicated point of contact.
