Microsoft Advertising published rules for AI-generated and synthetic content in ads

Microsoft Advertising published dedicated guidance covering AI-generated, AI-manipulated and other synthetic content in ads. Advertisers must disclose AI-created or materially altered creative where disclosure is required, preserve provenance signals, and remain liable for deceptive creative whether or not it is labelled.

The guidance sets out three things. Advertisers should preserve watermarks, metadata and other provenance information identifying how AI-generated content was created rather than stripping it during production. Where disclosure is required, the advertiser is responsible for making it clear and placing it close to the relevant content, and Microsoft recommends embedding the disclosure directly into image and video assets instead of relying on ad copy. And a label does not cure a deceptive ad: creative containing prohibited deepfakes, impersonating people or organisations, or using someone's likeness or voice without the required authorisation stays prohibited and subject to removal. Compliance with local law on disclosures, permissions and likeness rights remains the advertiser's responsibility in every market a campaign runs in.

Microsoft's guidance lives at help.ads.microsoft.com/apex/index/3/en/60265.