Can I use AI-generated product videos for advertising? A 2026 platform-policy and disclosure checklist for ecommerce brands
AI-generated product videos can run in ecommerce ads, but brands must verify product claims, rights, platform labels, and local disclosure rules before launch.
Lamina Team
Product Team @ Lamina
Can ecommerce brands run AI-generated product videos as ads?
Yes—provided the finished ad is truthful, cleared for commercial use, and disclosed when AI changes something a reasonable viewer would take as real.
How you made it is beside the point. If generated footage shows a skincare product delivering an unsubstantiated result, passes a synthetic actor off as a customer, or gives the product a feature it lacks, you have an ad problem—whether it runs on Facebook, Instagram, YouTube, or Search. Review what the viewer is left believing.
| Metric | Value | Source |
|---|---|---|
| Meta labeling threshold | Meta says AI labels can apply when images or videos are created or significantly edited with its generative-AI tools; insignificant edits without a photorealistic human may not receive a label. | about.fb.comas of 2025-02-03 |
| Meta tool coverage | Facebook and Instagram may automatically label ad content created or edited with certain Meta and third-party generative-AI tools, including tools such as Photoshop and DALL-E; availability can vary by region. | facebook.com |
| Google advertiser label option | Beginning in July 2026, Google allows advertisers to add text or visual AI labels in image and video creatives or use its AI-label setting across Google Ads, Display & Video 360, Campaign Manager 360, Merchant Center, and Ads Editor. | support.google.comas of 2026-07-09 |
| EU deepfake disclosure date | From 2026-08-02, EU Article 50(4) disclosure duties apply to deployers of AI-generated or manipulated image, audio, or video that qualifies as a deepfake. | lewissilkin.comas of 2026-07-31 |
| Human-readable EU label expectation | A machine-readable marking alone is insufficient where people exposed to the content cannot immediately see a clear and distinguishable disclosure. | lewissilkin.comas of 2026-07-31 |
| Jurisdictions Google flags | Google identifies the EU, India, and New York as jurisdictions with regulations requiring disclosures and/or labels for certain ads using AI-generated or edited assets. | support.google.comas of 2026-07-09 |
What needs disclosing in an AI-generated product video ad?
Disclose AI use where it materially shifts authenticity, identity, or representation in a way that could mislead the audience. Give close scrutiny to a realistic AI presenter; a cloned or synthetic voice; a made-up customer-style endorsement; an altered real person; or a product demonstration that appears real while being materially simulated.
A disclosure does not rescue an inaccurate ad. It cannot cure an implied testimonial that never occurred or a claim your business cannot substantiate. An aesthetic AI-generated background that leaves shoppers’ understanding of the item intact is less likely to trigger the EU deepfake duty in the legal guidance; altering the product itself carries a different risk.
“The digital industry is embracing AI in all of its splendor at breathtaking speed. We are certainly at a critical inflection point with generative AI,” said David Cohen, CEO, IAB. “While AI is transforming how we work from ideation to execution and measurement, we must get transparency and disclosure right, or we risk losing the trust that underpins the entire value exchange. We’re giving the ecosystem tools it needs to drive responsible innovation.”
How do Meta and Google treat AI-generated product video ads?
Meta and Google have AI-ad labeling tools. Your own review of claims, rights, and disclosures still matters. Meta says Facebook and Instagram can automatically label ads made or edited with eligible Meta and third-party generative-AI tools, and Google now allows advertisers to apply a label in eligible image and video workflows.
Across Google properties, the AI-label setting marks designated assets in My Ad Center’s “How this ad was made” area. Users reach it through the three-dot menu on Search, YouTube, and Discover ads. Google is explicit: its labeling option does not guarantee compliance with local law, so the campaign team still needs a market-by-market call before trafficking.
AI product-video ad pre-flight checklist
Classify the finished asset, not merely the workflow
Log whether AI handled minor production work only, or generated or significantly altered a realistic person, voice, place, event, product, or demonstration. Keep the final exported file, the prompt or generation log when available, and a brief note on what viewers may read as real.
Check product truthfulness, frame by frame
Match every visible product detail, use case, performance statement, before-and-after implication, and comparison against substantiation your business can produce. Fix generated details that make the item appear different, larger, more capable, or more effective than it actually is.
Keep a synthetic presenter distinct from a real customer
A synthetic person can deliver a brand-approved script or demonstrate a feature you can prove. Do not present that person as a real buyer, customer reviewer, independent expert, or someone who received a result that never happened.
Clear commercial rights and consents
Verify the AI provider’s commercial-use terms. Retain releases for real people, voices, digital replicas, product photography, music, and third-party inputs. Generated output does not erase the need to clear the source material or the likenesses shown.
Use the platform label control that is available
For Meta, preview the finished ad and check the ad-information experience before publishing. In Google campaigns, designate AI-generated or AI-edited assets in the available AI-label setting, or place a clear label in the creative where that fits the campaign decision.
Add a visible disclosure when the audience and law call for one
For EU delivery, assess whether the creative is deepfake content covered by Article 50(4). Where disclosure is required, make it prominent and readable by a person. Metadata or an invisible provider mark is not enough.
Geo-check the media plan and retain approval evidence
Review every jurisdiction where the ad will run, including markets Google identifies as having relevant AI-ad disclosure rules. Store the final creative screenshot, label decision, substantiation, approvals, release files, and platform setup record together; you may need to review the campaign after it launches.
Do all ecommerce ads touched by AI need the same label?
No. Materiality is the workable standard: step up labeling and review when AI changes what consumers might believe about the ad’s identity, authenticity, or representation.
The IAB framework calls for a risk-based approach, not blanket labels for every AI use. That is the rule an ecommerce creative team can actually run: minor production assistance and a photorealistic synthetic customer story belong in different approval lanes. Before media spend starts, a human art director and legal or policy reviewer should inspect brand-critical hero assets.
What is the safest default for AI product video advertising?
Use platform disclosure controls where they exist, add a clear audience-facing disclosure for realistic or materially deceptive-looking synthetic content where applicable, and retain evidence behind each launch decision. That keeps automatic detection from being mistaken for a full compliance program.
AI generation is still a practical route to new product concepts, complex styling, on-model scenes, and material detail without a traditional shoot. The work sits in the brief and approval gate: state the real product facts, rule out invented testimonials and unsupported outcomes, then inspect the exported ad instead of trusting the generation prompt. This checklist is research guidance, not legal advice.
Continue reading
Benchmark: Can AI-generated Shorts and Reels produce on-brand ecommerce product video ads without distorting the product?
A 30-SKU benchmark design for testing whether AI Shorts and Reels preserve the exact product, plus measured generation time, cost, and release gates.
Lamina Team
Product Team @ Lamina
EU AI Act compliance workflow for ecommerce AI product images and video ads: a pre-publish checklist for labeling, provenance metadata, and disclosure decisions
A pre-publish EU AI Act workflow for ecommerce images and video ads: check provenance, assess deepfake risk, place disclosures, and retain evidence.
Lamina Team
Product Team @ Lamina
Benchmarking AI video generation for ecommerce ads: from product assets and a brand kit to short, on-brand product video variants
Benchmark ecommerce AI video with fixed product inputs, scored approval gates, controlled variants, and account-level ad outcomes—not a beauty contest.
Lamina Team
Product Team @ Lamina