Skip to content
Start a Project
tech@softlixx.com
+92 347 0970 627

AI ContentBy Softlixx TeamPublished Updated 11 min read

Can you use AI-generated video in ads, and what must you disclose?

I read what YouTube, TikTok and Meta require, what regulators in the US, EU and UK say, and what surveys show about how viewers react.

On this page

The short answer

Yes, you can use AI-generated video in ads. The YouTube and TikTok pages I read allow it and ask for a label when the video looks real or AI changed it beyond minor edits, and Meta labels ads made with AI. Platforms can also add a label without you. Regulators in the US, EU and UK add that AI does not change the rules: an ad still cannot mislead, and fake testimonials are banned. That is what I found after reading the sources below.

Key takeaways

  • YouTube and TikTok allow AI-generated video if you disclose it when it looks real or AI changed it beyond minor edits. Meta labels ads made with AI tools.
  • Platforms can label your ad without you. YouTube reads C2PA metadata, and TikTok says undisclosed AI ads are rejected or restricted.
  • In the US, the FTC bans fake testimonials, including from people who do not exist. AI avatars are not banned, but can still be deceptive.
  • In the EU, Article 50 of the AI Act applies from 2 August 2026: deepfakes must be disclosed, with fines up to €15 million or 3% of turnover.
  • In the UK there is no blanket duty to disclose AI in ads, but disclosure will not fix a misleading message.
  • Viewers want to know: 76% of Americans say telling AI content from human content is extremely or very important, and 53% doubt they can.

What can AI video actually do today?

Short, realistic clips with sound, and advertisers are already using it. I started with Google's own documentation for one leading model.

The Google Cloud documentation for Veo 3.1, last updated on 1 October 2026, calls it Google's latest line of video generation models. It lists video lengths of 4, 6 or 8 seconds, extending videos, sound generation, output up to 4K on its main model, and support for Content Credentials (C2PA). The same page gives this version a retirement date of 17 November 2026 or later, so the specifications move fast. Softlixx sells AI video production, so I have a stake in this topic.

Advertisers are moving fast too. In an IAB and Sonata Insights survey, 83% of 104 US ad executives said their company has deployed AI in the creative process, up from 60% in 2024, and 71% of 505 Gen Z and Millennial consumers believe they have seen an AI-made ad, up from 54%.

AI in advertising is growing on both sides of the screen

Ad executives who say their company has deployed AI in the creative process

2024 study60%
2026 study83%

Bars use a scale of 0 to 100%. 104 US ad executives in 2025 and 2026, 75 in 2024.

Gen Z and Millennial consumers who believe they have seen an AI-made ad

2024 study54%
2026 study71%

Bars use a scale of 0 to 100%. 505 US consumers in 2025 and 2026, 300 in 2024.

Source: IAB.

My readingOne clip is at most 8 seconds in this model, so I read a longer ad as several clips joined or extended. How long an ad should be is covered in the research on video ad length.

What do YouTube, TikTok and Meta require?

They allow AI video, and they want a label when it looks real or AI changed it a lot. Each one words it differently, so I read all three.

YouTube's help page says creators must disclose AI content that makes a real person appear to say or do something they did not do, alters footage of a real event or place, or generates a realistic scene that did not occur. Content that is not realistic, or edits that are minor, need no disclosure. Disclosing will not limit a video's audience or affect its eligibility to earn money. I treat this creators' rule as covering any video you upload.

Realistic AI content and meaningful changes require disclosure, while non-realistic or minor edits don't.

YouTube, Help page on disclosing use of GenAI content. Read the source

TikTok's advertising policy, last updated in April 2026, allows significantly edited media and AI-generated content if you apply the AIGC label or add a clear disclaimer, caption, watermark or sticker of your own. Significant means content that is completely AI-generated, a main subject shown doing something they did not actually do, or made to say something they did not say using AI voice-cloning. If TikTok identifies AI-generated content that has not been disclosed, the ad will be rejected or restricted. It also does not allow the misuse of a public figure's likeness without permission.

Meta's post on ads transparency, from February 2025, says it labels ads created or significantly edited with its own generative AI creative tools. When those tools add an AI-generated photorealistic human, the label appears next to the Sponsored label instead of behind the three-dot menu.

What each platform asks you to do, and what it does
PlatformWhat you doWhat the platform does
YouTubeAnswer Yes to the AI use question in YouTube Studio when the video is realistic or meaningfully AI-altered.Adds a label in the video player for photorealistic content, or in the expanded description for non-photorealistic or animated content. May label it itself.
TikTokApply the AIGC label, or add a clear disclaimer, caption, watermark or sticker, for content significantly modified by AI.Rejects or restricts the ad if it identifies AI-generated content that was not disclosed.
MetaThe post describes no separate advertiser step. Labels are applied by Meta.Adds an AI info label in the About this ad menu, or next to Sponsored when the ad includes an AI-generated photorealistic person.

Sources: YouTube Help, TikTok Advertising Policies, Meta.

My readingThe three companies word it differently, but I would use one test for all: does the video look real, or did AI change it beyond a minor edit? If yes, disclose. I read only YouTube, TikTok and Meta, not Snapchat, LinkedIn, X or Google Ads, so check those separately.

Can platforms tell a video is AI-made if I do not say?

Often, yes. Tools and platforms are being built to pass a signal along with the file, so a label can appear even if you did not add one.

YouTube says it may automatically label content made with its own GenAI tools, content that contains C2PA metadata, and content its internal systems detect as AI. A label triggered by C2PA metadata or manual review cannot be adjusted by the creator. The Veo 3.1 page lists support for C2PA, and the European Commission says providers must add machine-readable marks to enable the detection of AI-generated content. Meta said it would detect ads made with third-party AI tools through industry-standard signals.

My readingThe file may carry the signal, and the platform may apply the label anyway. I take that as a good reason to answer the disclosure question honestly.

What do the law and advertising regulators say?

The rules differ by country, and none of the three I read bans AI ads. They add duties: to label deepfakes in the EU, to avoid fake testimonials in the US, and to keep ads honest in the UK.

In the US, the FTC's rule on consumer reviews and testimonials has been in effect since 21 October 2024. Its announcement says it addresses reviews and testimonials that misrepresent that they are by someone who does not exist, such as AI-generated fake reviews. It prohibits creating, selling, buying or spreading them when the business knew or should have known they were fake or false.

Fake reviews not only waste people's time and money, but also pollute the marketplace and divert business away from honest competitors…

Lina M. Khan, then Chair of the US Federal Trade Commission. Read the source

The FTC's questions and answers take up AI avatars directly. The rule has no blanket prohibition on AI-generated avatars in marketing. A testimonial from one is prohibited only if the underlying testimonial was fake or false, and avatars could still be deceptive under the FTC Act. Using a celebrity avatar without the celebrity's permission would also violate the rule, if reasonable consumers would think the celebrity gave a testimonial. For how paid creator videos are licensed and disclosed, see the UGC research.

In the EU, the Commission's guidelines page, last updated on 6 August 2026, says Article 50 of the AI Act applies from 2 August 2026: providers must add machine-readable marks, and deployers must inform people when they are exposed to deepfakes. Its Quick Facts page defines a deepfake as AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful. Fines go up to €15 million or 3% of worldwide annual turnover, and marking has a grace period until December 2026 for generative AI systems placed on the market before 2 August 2026.

In the UK, the ASA and CAP said in their May 2025 guidance that there is no blanket legal requirement to disclose the use of AI in ads, because their codes have no AI-specific rules and the existing rules apply however content is made. Their June 2026 guidance adds that an AI-generated celebrity endorsement that looks real enough to mislead consumers is likely to breach the Code.

What three regulators say about AI in ads
WhereWhat it says
US (FTC)No blanket ban on AI avatars. Fake or false testimonials are banned, including from people who do not exist. A celebrity avatar needs permission.
EUDeepfakes must be disclosed, and AI providers must add machine-readable marks. Fines go up to €15 million or 3% of worldwide annual turnover.
UK (ASA and CAP)No blanket duty to disclose AI in ads. The normal rules on misleading ads apply, and the advertiser stays responsible for the output.

Sources: Federal Trade Commission (announcement), Federal Trade Commission (questions and answers), European Commission (guidelines), European Commission (Quick Facts), ASA and CAP (May 2025), ASA and CAP (June 2026).

My readingI am not a lawyer, and I read only the US, EU and UK rules, not Pakistan's. A realistic AI actor playing a customer could fit the EU's deepfake definition, but I cannot tell you whether it does. Ask a lawyer who knows the markets you sell in.

How do viewers react to AI ads, and does a label hurt?

Viewers are more wary than advertisers think, and the studies disagree about labels. Two surveys and one experiment are worth knowing.

In a Pew Research Center survey of 5,023 US adults in June 2025, 76% said it is extremely or very important to be able to tell whether pictures, videos and text were made by AI or by people. Yet 53% were not too or not at all confident they could detect it.

IAB, which describes itself as the leading trade organization for the digital advertising ecosystem, found a gap in expectations in its survey: 82% of ad executives believed Gen Z and Millennial consumers feel very or somewhat positive about AI-generated ads, against 45% of the consumers themselves.

Viewers want to know, and executives are more upbeat than consumers

Americans on telling AI-made content from human-made (Pew Research Center)

Say it is important to be able to tell76%
Not confident they can tell53%

Bars use a scale of 0 to 100%. 5,023 US adults, June 2025.

Feel positive about AI-generated ads (IAB and Sonata Insights, 2026)

Ad executives think consumers do82%
Consumers who actually do45%

Bars use a scale of 0 to 100%. 104 US ad executives and 505 US Gen Z and Millennial consumers.

Sources: Pew Research Center, IAB.

On disclosure, the same survey found that 73% of the consumers said knowing an ad was made with AI would increase or make no difference to their likelihood to buy. More than half wanted disclosure of an ad that is 100% AI-generated, uses AI video or has AI images.

When AI creates material risk that consumers will be misled about what they're seeing, hearing or interacting with, advertisers should disclose.

Debra Aho Williamson and Jack Koch, authors of the IAB and Sonata Insights report. Read the source

A peer-reviewed experiment tells a more mixed story. In a study in Frontiers in Computer Science, published on 21 August 2026, 286 participants saw a simulated product page with an AI-generated image labelled AI-generated, labelled human-generated or unlabelled. For a power bank, the AI label reduced perceived authenticity, esthetic appeal and social presence. Yet AI labels also carried a positive direct effect on purchase intention, and the overall effect looked weak or even negative in raw form.

My readingBoth have limits. The IAB survey covers 505 US Gen Z and Millennial consumers. The experiment used power banks and perfume, a young, highly educated, AI-familiar sample, and images, not video. Neither is a forecast for your audience. I take from them that viewers want to know, and that a label is not shown to be a deal-breaker.

How do I check an AI video ad before it runs?

Ask the two questions the UK advertising regulator suggests. It is the clearest test I found.

The ASA and CAP suggest asking whether the audience is likely to be misled if the use of AI is not disclosed and, if so, whether the disclosure clarifies the ad's message or contradicts it. Their example is an AI-generated image showing the effect of a cosmetic product that does not accurately reflect real-world results, with a disclosure that AI was used.

However, it is important to remember that disclosure alone is very unlikely to mitigate the harm caused by a fundamentally misleading message.

ASA and CAP, the UK advertising regulator and its code-writing body. Read the source

Responsibility stays with the advertiser. The June 2026 guidance says that if an AI tool produces something problematic, the advertiser remains responsible for any output used in their ads. It cites a ruling where an ad made with Google's automated tools was still the advertiser's responsibility.

What should a business do first?

If I were starting today, I would work in this order: decide if it looks real, use the platform's label, keep the ad honest, then check the rules where the audience lives. It is my reading of the sources above.

  • Ask if the video looks realIf a viewer could take a person, place or event for real, treat it as needing a label.
  • Use the platform's own labelAnswer YouTube's AI use question, apply TikTok's AIGC label, and expect Meta's automatic label.
  • Never pass AI off as a customerThe FTC bans fake testimonials, including from someone who does not exist.
  • Get permission for anyone's face or voiceTikTok and the FTC both say a public figure's likeness needs permission.
  • Keep every claim true without the labelDisclosure alone will not fix a misleading message, the UK regulator says.
  • Check the rules where your viewers liveThe US, EU and UK rules differ, and I did not read Pakistan's. Ask a lawyer about your markets.
  • Review the final ad yourselfThe advertiser stays responsible for AI output, so someone should watch the finished ad and read every claim.

Questions people ask

Do I have to label every AI-generated video ad?

No single rule covers every case. YouTube asks for disclosure of realistic or meaningfully AI-altered content, and TikTok of significantly AI-modified content, but neither for minor edits such as color or lighting. The UK has no blanket duty to disclose AI in ads, while the EU requires deepfakes to be disclosed. TikTok says that when unsure, it is best to label.

Can I use an AI avatar as a spokesperson?

Often, yes. The FTC says its rule has no blanket prohibition on AI-generated avatars in marketing. But an avatar giving a fake or false testimonial is prohibited, and avatars can be deceptive under the FTC Act. TikTok also bans misusing a public figure's likeness without permission.

What happens if I do not label an AI ad?

It depends on the platform. TikTok says an ad with undisclosed AI-generated content will be rejected or restricted. YouTube may label it itself, and can penalize creators who consistently do not disclose, up to removal of content or suspension from its Partner Program. Meta said it would detect ads made with third-party AI tools and label them.

Does labelling an ad as AI-made reduce sales?

The evidence I found is mixed. In an IAB survey, 73% of Gen Z and Millennial consumers said knowing an ad was made with AI would increase their likelihood to buy or make no difference. In a peer-reviewed experiment, AI labels reduced perceived authenticity but had a positive direct effect on purchase intention. Neither tested live video ads, so test on your own audience.

Sources

How I checked this. I opened every source linked here and checked each figure and each quotation against the source page, most recently on 3 October 2026. Anything in a box marked “My reading” is my own interpretation, not a finding of the studies. Studies change and new ones appear, so I will update this piece and change its dates when the evidence does.

KEEP READING

See all research