Microsoft Advertising AI Content Policy: Provenance Must Stay, Labels Left to Local Law

Search Engine Land reported on September 14 that Microsoft Advertising has published a Help Center page titled “AI-generated and synthetic content in advertising,” setting out what it expects from ads built or altered with AI tools. Microsoft’s page itself carries no publication date, no author line, and no effective date. Read past the headline framing, and the AI content policy is narrower than a labeling mandate: keep the provenance data, don’t fake a real person, and add a visible disclosure only where local law says you must.

What the Policy Page Says Advertisers Must Do

The policy page sits in Microsoft Advertising’s Help navigation. In a section titled “Your responsibilities as an advertiser,” it says advertisers must follow the laws and regulations that apply wherever an ad runs: obtaining rights, permissions or consent; providing disclosures when required; and preserving watermarks, metadata or other provenance information that identifies how content was made. Requirements may differ by market and by content type, the page says, so advertisers should check the rules for their own campaign and seek legal guidance when needed. Some AI-generated or AI-manipulated content is barred outright, deepfake content among it, regardless of whether it carries a disclosure. Advertisers must also comply with Microsoft Advertising’s separate Unacceptable content policy before running AI-touched creative, the page adds, without spelling out what that policy covers.

How Microsoft’s Own Tools Mark AI Content

A section called “How Microsoft supports transparency” describes what happens when Microsoft’s own generative tools make the asset. When Microsoft AI tools generate an image, audio or video, the page says, Microsoft applies machine-readable provenance information, metadata and imperceptible watermarks to help indicate the content was created or edited using AI. The page notes that this marking “may not appear as a visible label.” That’s why the next line carries the weight: advertisers remain responsible for adding any visible or audible disclosure required by applicable law, placed close to the content it describes. The page recommends embedding AI disclosures directly into image and video assets and says advertisers may also use the Disclaimer in ads feature when their ad format supports it, without describing what that feature does beyond naming it.

“A disclosure does not make deceptive content acceptable.”

— Microsoft Advertising Help, “AI-generated and synthetic content in advertising”

Microsoft Advertising may reject, restrict or remove ads that contain prohibited deepfake content or that use AI-generated content to mislead consumers, impersonate a person or organization, omit a required disclosure, use a person’s likeness or voice without required authorization, or remove or interfere with required machine-readable information, the page says.

What does Microsoft Advertising’s AI content policy require?

Microsoft Advertising’s policy page says advertisers using AI-generated, AI-manipulated or other synthetic content must follow the laws and regulations that apply wherever the ad runs, and must preserve any watermarks, metadata or other provenance information the content carries. Visible or audible disclosure is required where applicable law demands it, placed close to the content it describes; the page states no Microsoft label requirement of its own and names no specific markets. Deepfake content is prohibited regardless of disclosure, and Microsoft says a disclosure does not make otherwise deceptive content acceptable.

How the Coverage, Google and the IAB Compare

The Search Engine Land write-up says Microsoft Advertising “has published dedicated guidance covering AI-generated, AI-manipulated and other synthetic content used in ads.” It doesn’t quote Microsoft directly and gives no date for the page either. Its paraphrase lines up with the page: preserve watermarks and metadata, keep required disclosures clear and close to the content, and know that labeling deceptive creative “AI-generated” won’t make it acceptable.

Microsoft’s approach reads differently from Google’s. Google’s July 9 policy note says the company began permitting text or visual AI labels directly within image and video ad creatives that were generated or modified using AI, “to help advertisers comply with emerging AI transparency regulations”; it names AI regulations in the European Union, India and New York as requiring such disclosures, and adds that using its AI label setting “doesn’t guarantee compliance with specific regulations.” That caveat mirrors Microsoft’s own “seek legal guidance when needed.” Google’s July 9 blog post on the change also says the company already embeds “imperceptible signals, like SynthID,” into outputs from its generative AI tools. But Microsoft’s page names no jurisdictions and describes no labeling toggle of its own; the closest tool it names is the Disclaimer in ads feature.

The gap between disclosing and labeling also runs through the IAB’s risk-based AI disclosure framework, which recommends labels for prompt-generated images and video, synthetic avatars, certain digital twins and some synthetic voices, but exempts ad copy and routine retouching outright. Microsoft’s trigger phrase is “generated or materially changed using AI”; the page names no exemption list. The EU AI Act’s Article 50 transparency duty requires providers of generative AI systems to mark their outputs in a machine-readable format and requires deployers to disclose deepfakes, but it stops short of a blanket duty to label ad copy. The table below lines up what each source states.

Before the next upload, a media buyer running AI-touched creative on Microsoft Advertising should keep whatever provenance data, metadata or watermark ships with the asset intact, add a visible disclosure wherever the target market’s law requires one, and leave deepfake content out of the mix regardless of disclosure.