AI Content Labels in Asset Studio: What D2C Advertisers Need to Know

If your D2C team has been using AI to generate or touch up product images, lifestyle shots, or video creative for Google Ads, there's a new setting sitting in your account that you're either already using, legally required to use, or one regulatory update away from needing. Starting July 9, 2026, Google rolled out AI content labels inside Asset Studio, giving advertisers a built-in way to disclose when an image or video asset was generated or modified using AI, and in some markets, that disclosure isn't optional anymore.

The short version: Google now lets you label AI-generated or AI-edited assets directly in Asset Studio, the labels satisfy Google's own policy against text overlays and watermarks, and they exist specifically because the EU, India, and New York have all introduced AI transparency rules that apply to advertising. Whether you need to use the setting depends on where your ads run and what your creative actually contains, but every D2C advertiser running AI-assisted creative needs to understand the mechanics now, not after a compliance question lands on your desk.

What Exactly Are AI Content Labels in Asset Studio?

The AI label setting lets you identify an individual image or video asset as created or edited with AI, directly inside Asset Studio. It's rolling out gradually across five platforms: Google Ads, Display & Video 360, Campaign Manager 360, Merchant Center, and Google Ads Editor.

Mechanically, it works like this:

• Go to Tools > Asset Studio > Asset library.

• Select an asset, then choose "Label this asset as created or edited with AI" or "Don't label this asset."

• You can add an "AI label" status column to the Asset library and asset reports to see labeling status at a glance.

• In some geographies, the label also appears directly on the ad itself as a visible disclosure, not just in backend reporting.

• Once Google applies a label automatically, either on your behalf or because a region requires it, you cannot overwrite it yourself.

Separately, Google added a consumer-facing "How this ad was made" panel inside My Ad Center on Search, YouTube, and Discover, so people viewing your ads can check whether generative AI was used to create or alter what they're looking at.

Why Did Google Build This Now?

This isn't Google getting ahead of a hypothetical concern, it's responding to law that's already in force. Three regulatory regimes converged in mid-2026:

• New York's Synthetic Performer Act took effect June 9, 2026, exactly 180 days after being signed into law. It requires advertisers to conspicuously disclose when an ad features a digitally created human that isn't a real person, and the obligation applies to any campaign reaching New York consumers regardless of where the advertiser is based.

• India's amended intermediary rules took effect February 20, 2026, adding AI disclosure requirements on that side of the regulatory map.

• Article 50 of the EU AI Act became applicable August 2, 2026, requiring disclosure for realistic AI-generated or AI-manipulated depictions of people, objects, places, or events that could falsely appear authentic, and for AI-generated or AI-manipulated text used to inform the public on matters of public interest, though meaningful human review and editorial responsibility can eliminate the need for disclosure in some cases.
  

Google's own policy language is direct about this: these labels exist to help advertisers comply with emerging AI transparency regulations. Using the setting doesn't automatically guarantee compliance with any specific law, but it gives you the technical mechanism the regulations assume exists.

Do You Actually Have to Label Every AI-Generated Asset?

No, and this is the detail that trips up a lot of advertisers. Not every AI-generated or AI-edited asset needs a label. Requirements depend on three things: your content, your geography, and how AI was used to create or modify the asset. One industry commentator summed up the practical reality bluntly: most advertisers won't need this setting unless they're legally required to use it in a specific market.

That said, a few things are worth knowing regardless of whether you're currently required to label anything:

• Google may label some assets automatically. Content created with Google's own AI tools may receive an AI label without you choosing it, and once applied, that label can't be removed.

• Google's AI-generated images already carry invisible technical markers. Every AI-generated image and video from Google's own tools carries a SynthID watermark and C2PA metadata, invisible to the eye but machine-readable, regardless of whether a visible label is applied.

• The label doesn't replace your other disclosure obligations. Google is explicit that this update doesn't change the existing requirement that advertisers clearly identify who they are.

Which Markets Currently Require AI Disclosure in Ads?

These regimes were drafted independently and define their subject matter differently, so a single global campaign using a synthetic presenter can face three separate, non-identical disclosure standards depending on where it runs. There is no single label configuration that automatically satisfies all of them, legal review by market is still the safer path for anything involving a synthetic human performer specifically.

What Should D2C Advertisers Actually Do?

1. Check whether the AI label setting has appeared in your account. The rollout is gradual and staggered by platform and account, so don't assume you have it, or don't have it, without checking Asset Studio directly.

2. Inventory which creative assets were AI-generated or AI-edited. This includes product photography touch-ups,  AI-generated lifestyle imagery, and any synthetic voice or on-screen presence in video ads, not just fully AI-generated content from scratch.

3. Map your campaigns against the New York, EU, and India requirements specifically. If any campaign targets or is reasonably likely to reach consumers in those markets, treat disclosure as a compliance question, not a creative preference.

4. Get legal guidance for anything involving a synthetic performer. Because the three regimes define "synthetic" content differently and carry different penalty structures, a D2C brand running the same video creative across all three markets needs a compliance answer for each, not one blanket assumption.

5. Build labeling into your creative QA process going forward, not as an afterthought. Google now prompts advertisers to "review assets" during campaign creation when AI labeling may be required, treat that prompt as a checkpoint, not something to dismiss automatically.

6. Don't rely on the Google setting alone as proof of compliance. Google states plainly that using the AI label setting doesn't guarantee compliance with any particular regulation; it's the mechanism, not the legal determination.

How Does This Affect Everyday D2C Creative Testing?

For most D2C brands, the realistic day-to-day impact is smaller than the regulatory headlines suggest, a lot of AI-assisted product creative (background removal, minor touch-ups, resizing for different placements) sits well outside what New York, the EU, or India are targeting, which is aimed more squarely at realistic synthetic humans and content that could be mistaken for authentic, unaltered reality. The bigger shift is procedural: creative teams generating high volumes of AI variants for testing now need a lightweight review step asking "does this need a label" before assets go live, especially for any video featuring an AI-generated presenter or spokesperson-style content.

For D2C teams running AI-generated creative at scale across multiple ad platforms and markets, this is exactly the kind of compliance-adjacent detail that's easy to lose track of when creative production and campaign management sit in different tools. A platform like NYX, which handles creative generation and campaign deployment together, is useful here because labeling and compliance checks can sit inside the same workflow that produces and launches the creative, rather than becoming a separate manual audit after assets are already live.

A Simple Example (Illustrative Scenario)

A D2C apparel brand running video ads across the US, UK, and several EU markets had been using an AI-generated spokesperson-style presenter in a handful of Performance Max video assets. After the July 2026 rollout, the brand's paid media lead audited existing video assets against the "How this ad was made" and Asset Studio settings, discovered two videos using a fully synthetic on-screen presenter were running against New York and EU audiences, and applied labels through Asset Studio ahead of the August 2 EU deadline while consulting legal on whether the New York disclosure specifically applied. 

The fix took under a day once the audit identified the affected assets, the harder part was finding them in the first place, since the videos had been created months earlier by a different team member with no record of which assets used AI-generated presenters.