Ad policies and review

Do I have to tick Meta's AI disclosure box for B2B ads made with ChatGPT or Midjourney images, and does it hurt delivery?

By The Tofu team ·

Not for most B2B ads. Meta makes AI disclosure mandatory only for ads about social issues, elections or politics with realistic AI media. For other ads, Meta's help says that where local rules require AI transparency (the European Region, California, New York, India and Taiwan), advertisers "in some cases" can self-disclose, and the label then shows next to Sponsored. Since June 1, 2026, Meta also labels ads it detects were made with third-party AI tools, using signals such as C2PA, whether you tick the box or not. Meta doesn't say the label affects delivery.

Ask Tofu next

Pick a question and Tofu Ads Agent answers it in chat.

As of November 2026.

What exactly does Meta require?

Only one hard rule. Meta's policy on ads about social issues, elections or politics says advertisers "must also disclose" when such an ad has a photorealistic image or video, or realistic-sounding audio, "created or edited using third-party generative AI tools" to "depict a real person as saying or doing something they did not say or do," to show a realistic person or event that doesn't exist, or to show a realistic event that isn't a true recording. If Meta finds you didn't disclose, "we will reject the ad," and "repeated failure to disclose may result in penalties against the advertiser."

For everyday B2B ads (a product launch, a webinar, a demo offer), Meta's help center doesn't say disclosure is required. It says: "You may choose to deliver ads to states or countries where local laws or guidelines include specific AI transparency requirements. In some cases, advertisers will have the opportunity to self-disclose that GenAI was used." It names "the European Region, California, New York, India and Taiwan." It doesn't explain what those local rules require, so ask your legal team if you target them.

A B2B ad can count as a social issue ad more easily than you'd think, for example if it talks about regulation or the economy. See why Meta flags B2B ads as social issue ads.

Where is the AI disclosure checkbox in Ads Manager?

The Meta help pages and policies we read don't name it or say where it sits. Jon Loomer, who writes about Meta advertising, wrote on June 18, 2026 that he found it "conspicuously hidden above primary text when creating an ad," and that it's "very easy to miss." Meta's help adds that "before publishing, you will see a message during the ad creation process letting you know when an AI info label will appear on your ad." Not every account may see the box, since Meta says self-disclosure is offered "in some cases."

What label does it add, and where do people see it?

The label is called AI info. Meta's help says that if you self-disclose, "the label will appear on the ad next to the Sponsored label." That's on the face of the ad, where people actually see it.

When Meta labels an ad on its own, the label usually sits one step away: "navigating to the three-dot menu, selecting About this ad, and then tapping on the AI info label." Meta's AI help page adds that "sometimes an AI info label will also appear next to the Sponsored label at the top of the ad," without saying when. For ads made with Meta's own tools, its 2025 announcement says a photorealistic AI-made person moves the label next to Sponsored.

Two more details from Meta's help: in some regions the "Sponsored" label now reads "Ad," and "depending on your region, the AI info label name or the Sponsored label name may be different." Jon Loomer says the label is much easier to notice when you tick the box than when Meta adds it.

Will Meta label my ChatGPT or Midjourney image even if I don't tick the box?

It can. Meta's help says: "We will automatically apply an AI info label when we detect your ad was created or edited by third-party generative AI tools." It explains that "Meta uses industry-standard detection methods, such as C2PA," and "when we detect this metadata, we label the content accordingly."

The start date comes from Meta itself. Its policy on social issue, election and political ads says that "beginning June 1, 2026, Meta will also use automated detection technology" to find ad media made or edited with third-party AI tools, and that "no advertiser action is required." Meta's newsroom post on AI in ads, updated June 1, 2026, says Meta will "begin automatically detecting ads created or edited using third-party AI tools through industry-standard signals," and doesn't limit that to political ads. Its post on the 2026 US midterms, updated the same day, says Meta is "adding more information to ads we detect have been created with third-party AI tools."

Whether a given ChatGPT or Midjourney file triggers the label depends on the signals, such as C2PA metadata, it still carries when you upload it. We couldn't confirm from OpenAI's or Midjourney's own help pages which signals each tool writes today (both blocked our page reader), and none of the Meta pages we read list the tools it detects. Don't assume an image will or won't be labeled. Also, Meta's AI help page says that "not all ads will have AI info, like when there are minor changes or enhancements like image resizing or color correction."

Does the AI info label hurt delivery?

Meta doesn't say. None of the Meta help pages, policies or announcements we read mention delivery, ranking, reach or cost for labeled ads. Meta describes the label as transparency information, grouped in About this ad with "payer and beneficiary details" and other ad details. Jon Loomer's article doesn't comment on performance either.

Any claim that the label cuts or doesn't cut results is a guess until Meta publishes something. If it matters to you, test it in your own account: run the same offer with an AI-made image and a photo or design made without AI, in an A/B test, and compare cost per lead.

Can an undisclosed AI image get my ad rejected?

For social issue, election and political ads, yes, as quoted above. For other ads, Meta's pages don't list a missing self-disclosure as a reason to reject. An AI image still has to follow the rest of Meta's rules, though. Meta's misinformation policy says it may "reject content submitted as an advertisement" when it is a photorealistic image or video, or realistic audio, "that was digitally created or altered" and "creates a particularly high risk of materially deceiving the public on a matter of public importance."

For B2B, the bigger risk is usually ordinary policy problems in the ad, not the AI. See why Meta rejects ads that ran fine for months and why Meta rejects B2B copy for personal attributes.

What about images made with Meta's own AI tools?

Meta labels those itself. Its help says it will "automatically apply an AI info label when you use features such as Background Generation, Image Generation, or Add Animation to create or significantly edit an image or video." In Taiwan and South Korea, it may also label ads when it detects AI-made photorealistic people created with Meta's tools. Meta's Ad Creative Generative AI Terms, which cover those tools, also say "you must not misrepresent Output as human-generated when it is not" and "you must not alter or remove any watermark or content-authenticating metadata." Meta's terms pages say these terms fold into its Self-Serve Ad Terms on November 20, 2026, but the UK competition regulator says Meta put that update on hold on October 28 with no new date, so for now they stand on their own. To switch those features off, and for the latest on the terms, see how to turn off Meta's AI edits to your ads.

Should a B2B advertiser tick the box anyway?

It depends on where the ad runs and what the image shows. A practical rule:

Your situationWhat to do
Social issue, election or political ad with realistic AI mediaDisclose. Meta requires it and rejects the ad if you don't
Ad served in the European Region, California, New York, India or Taiwan with AI-made or AI-edited imagesTick the box if Meta offers it, and check the local rules with legal. Jon Loomer says that in places with disclosure laws "it would appear that the checkbox is necessary"
Ad served only elsewhereMeta doesn't require it. Meta may still label the ad if it detects AI metadata
Image only resized or color-correctedMeta says minor edits may not get AI info

Jon Loomer's own advice is to "play it very safe" and tick the box when you use AI to make or change creative, "until we know more." The trade-off is a more visible label on your ad.

Sources

Related answers

More on

Let an agent build your next campaign.

Tofu Ads Agent reads your CRM, plans the funnel, builds the audience and creative, and stages full campaigns across LinkedIn, Google and Meta for your approval.

Try it free Get a demo