Quick answer
Instagram adds an AI info label when Meta detects industry standard AI image indicators, such as the AI information in the C2PA and IPTC metadata standards, or when you disclose AI use as you post. TikTok, YouTube and Pinterest also read file metadata, and TikTok and YouTube require you to disclose realistic AI content yourself. For a seller, the rule that matters most is older than any label: product images in an ad must be truthful and non-deceptive.
Nation, an AI store builder, publishes this page. It is open: describe what you make in the app onboarding and the store builds itself, and it does not decide how any platform labels content. See where Nation fits.
What the Instagram AI label is and what triggers it
The Instagram AI label is the AI info tag that Meta shows on posts it believes were made or changed with AI. The same system runs on Facebook and Threads. Meta started labelling organic posts in May 2024 under the name Made with AI, then renamed it AI info, a label people can click for more information.
Two things trigger the label. The first is metadata inside the file. Meta reads the AI information in the C2PA and IPTC technical standards, which lets it label images from Google, OpenAI, Microsoft, Adobe, Midjourney, and Shutterstock when those tools write that data. The second is you: if you disclose AI use when you post, Meta adds the label.
Where the label sits depends on how much AI was used. For content Meta detects was only edited with AI, the label moves to the post’s menu instead of showing on the post. Content Meta detects was generated by AI keeps the visible label. Images made with Meta's own AI carry a separate label, Imagined with AI, and content with a high risk of deceiving the public on an important matter can get a more prominent label.
- Must you disclose on Instagram and Facebook? Meta requires disclosure for photorealistic video or realistic-sounding audio that was digitally created or altered, and it may apply penalties if you do not.
- Still images are not on that required list, but an AI image can still be labelled automatically if its file carries the metadata.
- Scale: from October 1, 2024 through October 29, 2024, Instagram had over 330 million labelled pieces of content that people saw with the label.
AI labels on Instagram, Facebook, TikTok, YouTube and Pinterest compared
All four platforms below read provenance metadata in files. What differs is the label wording, what you are required to disclose, and whether you can remove a label that was added for you.
C2PA is the Coalition for Content Provenance and Authenticity. It publishes an open technical standard called Content Credentials: a record of where a file came from and how it was edited. IPTC is a photo metadata standard, and Meta reads the AI information in both. Meta names Google, OpenAI, Microsoft, Adobe, Midjourney, and Shutterstock as companies whose images it can label as they add this metadata.
TikTok AI-generated content label
TikTok has two labels. You can switch on the AI-generated content setting before you post, which shows creator labeled as AI-generated. TikTok also adds its own AI-generated label when an upload has Content Credentials attached, or when you used TikTok effects that rely on AI.
TikTok requires a label on all AI content with realistic images, audio, and video. It counts a real photo or video as significantly edited when it was modified by AI beyond minor corrections or enhancements. Once TikTok applies its auto label, you won't be able to remove the label. TikTok says the setting won't affect the distribution of a post that follows its rules, and it may remove content that leaves out a label it should have had.
Some AI content is banned even with a label, including AI content that falsely shows public figures making an endorsement, or being endorsed. A seller should never use an AI version of a real person to promote a product.
YouTube AI label and the AI use setting
YouTube requires creators to disclose when AI is used to meaningfully alter or generate photorealistic content. You do this during upload in YouTube Studio, under the AI use setting. YouTube's own list of examples that need disclosure includes AI generated extra footage of a real place for a promotional video.
You do not need to disclose non-realistic content that’s made with AI, or edits to realistic content that are minor. YouTube also exempts generative AI used for a video's outline, script, thumbnail, title, or infographic. A product thumbnail made with AI does not need the label on its own.
YouTube can add the label itself, including when its internal systems detect AI-generated or altered content. A label added because the file has C2PA metadata cannot be changed by the creator. Disclosure won’t limit a video’s audience or its earnings, but creators who keep skipping it risk removal of content or suspension from the YouTube Partner Program.
Pinterest AI label and AI labels on ads
Pinterest shows an AI modified label on Pins it finds were made or changed with AI. It reads metadata under the IPTC Metadata Standard and also uses classifiers that can flag AI images without metadata. If you think a label is wrong, Pinterest says to contact our support team. On promoted Pins, the label is not on the Pin itself; viewers find it behind the ellipsis icon.
Meta handles ads in a similar way. AI labels on Facebook and Instagram ads appear in About this ad, in the ad's three-dot menu. Meta detects ads made with outside AI tools through industry-standard signals. If Meta's own ad tools add an AI-generated photorealistic person, the label appears next to the Sponsored label. If Meta's tools make only small changes and add no photorealistic person, Meta will not apply any AI labels.
What a seller posting AI-edited product images must disclose
Platform labels describe how a file was made. The law cares whether the ad is honest. Under the FTC Act, advertising must be truthful and non-deceptive, and an ad can be deceptive when it omits information. Example: an AI-generated lifestyle photo of a real candle on a styled table is a staging choice; an AI image that makes the candle larger, glossier or a different colour than the one you ship misrepresents the product.
Show the product as it is
Use AI for backgrounds, lighting and scenes, not to change the product's size, colour, finish or features. If the image would mislead a buyer about what arrives, no label fixes that.
Disclose realistic AI people and scenes
An AI model wearing your jewellery or an AI video of your product in use is realistic AI content. Switch on the platform's AI setting: TikTok requires it for realistic images, audio, and video, and Meta for photorealistic video or realistic-sounding audio that was digitally created or altered.
Expect metadata to speak for you
Files exported from some AI tools carry Content Credentials or IPTC data. Platforms read it, and a label added that way on TikTok or YouTube can be permanent. Assume the label will appear and decide whether that is acceptable before you post.
Keep the original photos
Save the unedited product photos and the AI files. They show what you changed if a buyer, a platform or a regulator asks.
Check state rules
The FTC notes that every state has consumer protection laws that govern ads running in that state. This page is general information, not legal advice; rules vary by state, so check with a lawyer for a specific case.
Where Nation fits
Nation
Category: an AI store builder
Nation says one description of what you make builds the store, then agents handle orders, customers and delivery and run ads, content and social posts, with you approving the work. Nation does not decide how social platforms label AI content.
Limits today
- Status: open: describe what you make in the app onboarding and the store builds itself.
- Pricing and fees: from $39 a month plus a 2% fee per sale, with a 14-day trial for $1.
- Shopper marketplace: no shopper marketplace has been announced.
- Track record: no independent reviews or public stores yet.
Nation publishes this page. The labelling rules described here belong to each platform.
Describe what you makeQuestions about AI content labels
Why did Instagram put an AI info label on my post?
Instagram puts an AI info label on a post when Meta detects industry standard AI image indicators or when you disclose AI use. Most often the image file carried AI metadata in the C2PA and IPTC standards, written by the tool you used to create or edit it. Content Meta finds was only edited with AI has the label in the post’s menu instead.
Do I have to label AI images on Instagram and Facebook?
Meta's labelling rule for Instagram and Facebook requires disclosure for photorealistic video or realistic-sounding audio that was digitally created or altered, and Meta may apply penalties when people skip it. AI still images are not on that required list, but Meta can label them automatically when their files carry AI metadata, so the label may appear anyway.
What is the TikTok AI-generated content label?
The TikTok AI-generated content label tells viewers a post was made or significantly changed with AI. Creators add creator labeled as AI-generated themselves, and TikTok adds its own label when uploads have Content Credentials attached. TikTok requires a label on AI content with realistic images, audio, and video, and an auto label cannot be removed.
Does the YouTube AI label hurt views or monetization?
The YouTube AI label does not reduce reach on its own: YouTube says disclosure won’t limit a video’s audience or its eligibility to earn money. Skipping disclosure is the risk. Creators who repeatedly fail to disclose realistic AI content can face removal of content or suspension from the YouTube Partner Program.
Can I remove an AI label from a post?
Removing an AI label depends on the platform. On TikTok, once the auto label is applied you won't be able to remove the label. On YouTube, labels tied to C2PA metadata cannot be changed. On Pinterest you can appeal: contact our support team. Meta shows edit-only labels in the post’s menu.
Do AI-edited product photos need a disclosure in ads?
AI-edited product photos in ads need to be honest before anything else: under the FTC Act, advertising must be truthful and non-deceptive. Meta may add an AI label to ads made with AI tools, shown in About this ad. An AI image that changes how the product looks can mislead buyers, label or not.