Microsoft Advertising has published dedicated guidance for AI-generated, AI-manipulated, and other synthetic content in ads. The platform isn’t banning AI creative. It’s asking advertisers to treat transparency as seriously as targeting and budget.
What the new guidance says
The core requirement is accountability. Advertisers remain responsible for following applicable laws wherever campaigns run, including rules around disclosures, permissions, and the use of a person’s likeness or voice.
Microsoft wants advertisers to preserve watermarks, metadata, and other provenance information that shows how AI-generated content was created. When a disclosure is required, it needs to be clear and placed close to the relevant creative. Microsoft recommends embedding AI disclosures directly into images and video assets, though its existing ad disclaimer feature can also be used with supported formats.
But a label won’t fix bad creative. Microsoft says ads can still be rejected, restricted, or removed if they contain prohibited deepfakes, impersonate people or organizations, use someone’s likeness or voice without authorization, omit required disclosures, or interfere with machine-readable provenance information.
Why this matters for performance teams
AI creative is already standard for testing angles, refreshing ad fatigue, and scaling D2C campaigns. But machine speed doesn’t remove platform risk. A rejected ad can stall a promising test, and a policy violation can put an account in a tougher review queue.
There’s also a consumer trust angle. Microsoft is drawing a line between using AI to produce an ad and using AI to mislead the person seeing it. For growth marketers, that line affects creative strategy, not just compliance.
Your AI ad checklist
Before you submit AI-assisted creative, run through these checks:
- Review accuracy: Check people, products, places, claims, and events depicted in the creative.
- Keep provenance intact: Don’t strip metadata, watermarks, or machine-readable signals from AI assets.
- Add visible or audible disclosures: Where required, place them close to the content; embedded disclosures are preferred.
- Check every market: Disclosure rules vary, so verify requirements wherever the campaign will run.
- Don’t rely on “AI-generated” labels: Deceptive creative can still be rejected, restricted, or removed.
Microsoft’s own AI tools can add machine-readable provenance data, metadata, and imperceptible watermarks. Those signals are not necessarily visible to consumers, so advertisers may still need their own clear disclosures.
What to do now
Treat AI transparency as a creative QA layer, not a legal afterthought. Add a quick review step for synthetic assets, keep original provenance data, and build disclosure requirements into your creative briefs. That way AI can keep accelerating production without creating policy risk.
Source: Search Engine Land
