Where Pinterest's AI modified label appears

The bottom-left corner of a Pin is where most brands put their logo. It is the least busy part of a 2:3 crop, it survives the grid thumbnail, and it has been the default lockup position for a decade. On Pinterest that corner is now shared real estate: per Pinterest's own help documentation on Gen AI labels, the "AI modified" label sits bottom-left in close-up view.

That is a layout constraint, not a policy problem. And it behaves differently on organic Pins than it does on ads, which means the same creative needs two different plans depending on where it runs.

Two surfaces, two placements

Pinterest surfaces the same label in two structurally different ways.

On an organic Pin, the label renders in the bottom-left of the close-up view — the expanded view a user lands on after tapping through from the grid. It is a visible element drawn over or beside your creative, in a corner you did not allocate to it.

On an ad, the label is surfaced through the ellipsis menu. The creative itself stays clean; a viewer has to open the overflow menu to find the disclosure.

The asymmetry is the whole design problem. On organic Pins you have a collision to avoid. On ads you have the opposite issue: a disclosure that exists but that a viewer has to open a menu to find, which matters if you were relying on the platform label to carry a disclosure obligation you actually own.

Design as if it always renders

Pinterest applies the label from three triggers — IPTC metadata, owner disclosure, and its own classifiers, which it says catch generated content "even if the content doesn't have obvious markers." Only the middle one is yours to withhold, and withholding it does not stop the other two.

For a designer that collapses to one premise: if any part of the image was generated or modified with AI, build the template as though the label will appear. Trying to engineer a clean corner by stripping the file is a losing trade for reasons that have nothing to do with layout — see why C2PA manifests die on upload.

Pinterest does not publish the label's dimensions, and platform chrome changes. So what follows is a reserved band, not a measured box, and you confirm the real footprint on your own Pin before committing a template to it.

The wrinkle that makes this a template problem rather than a one-off fix: the grid and the close-up are different frames of the same file. In the grid, a Pin is a small thumbnail competing with a wall of other thumbnails. In close-up, it is the whole screen with platform chrome laid over it. The label appears only in the second. So a bottom-left logo looks correct everywhere you normally preview — grid, design file, phone photo roll — and collides only on the surface you were not checking.

The rule that resolves both frames at once is to load the top of the image and leave the bottom quiet. Top-weighted layouts survive grid downscaling better anyway, because the eye lands high in a small thumbnail, and they leave the close-up's bottom band available for chrome you do not control. One change, two problems.

A layout that avoids the collision

Run through a typical commercial Pin and count how much of it clusters at the bottom:

• Logo or wordmark lockup
• Domain or URL strip
• Price or offer sticker
• CTA button ("Shop now", "Get the recipe")
• Step counters on multi-step how-to Pins
• Legal or asterisk copy on anything regulated

Most of that is bottom-anchored by habit rather than by necessity. The 2:3 canvas has plenty of usable vertical space and the bottom band is simply where designers have been trained to park furniture. Moving it costs nothing except a template revision. Six changes make the template label-proof:

1. Reserve the bottom band of the close-up frame. Treat the lower strip of the image as platform chrome, not creative space. Nothing you need a viewer to read goes there.
2. Move the brand lockup to the top. Top-left or top-right both work and both survive the grid crop. This is the single change that fixes most templates.
3. Keep the headline in the upper two-thirds. Pinterest surfaces text-on-image at thumbnail size anyway, and the top of the frame is what the grid shows first.
4. Keep the bottom band visually quiet. If the label lands on a high-detail, high-contrast area, it becomes hard to read and the viewer stops to squint at it. A calm gradient or flat area under the label costs you nothing and keeps attention on the creative.
5. Never put load-bearing information in the bottom corner. Price, discount code, dates, dosage, sizing, disclaimers. If a viewer has to read it to act, it does not go where a label can land on it.
6. Put the CTA above the reserved band, not flush to the bottom edge.

The zone rule, in one table:

Zone	Belongs here	Must not go here
Top third	Logo, headline, category cue	—
Middle third	Product, subject, hero imagery	—
Bottom band	Texture, gradient, negative space	Logo, price, CTA, legal copy, step numbers

If you build Pins from generated imagery, the cleanest way to enforce this is at generation time rather than in post: prompt for compositions with open, low-detail space along the bottom edge, then place type over it. A text-to-image prompt like "product centred in upper two-thirds, soft out-of-focus surface across the lower edge, even lighting, no text" gives you a frame that takes a label without fighting it. Composite the type in a design tool so the type layer stays editable when the template changes again.

Ads are the inverse problem

Because the ad label lives behind the ellipsis, your ad creative arrives visually clean. That is convenient and it is also a trap.

A label a viewer has to hunt for does not discharge a disclosure obligation you hold independently — under an advertising code, under a brand policy, or under a regulation that applies to your market. The platform's placement decision and your compliance position are separate questions, and the ellipsis menu answers only the first. If you have an obligation, the disclosure belongs in copy you control: the ad description, an on-creative line you placed yourself, or a landing page. AI ad disclosure compliance covers the shape of those obligations, and three EU AI disclosure icons explained covers the machine-readable side.

The practical version: for paid Pins, write your disclosure line into the description field as a default and stop thinking about it. It costs one sentence and removes a category of argument later.

Check it on the real surface, not in the design file

Every template change gets one verification pass, and it takes about two minutes:

1. Publish the Pin to a private or low-traffic board.
2. Open it and tap through to close-up view. The grid thumbnail is not the surface you are testing.
3. Do it on mobile and on desktop. Chrome placement is not always identical.
4. Screenshot both. Compare against your design file and look specifically for overlap with type, logo, or product edge.
5. If a label appears where you did not expect one, that is useful information about how Pinterest classified the asset — note it against the creative.

Fold this into whatever pre-publish routine you already run for Pinterest video and Pin formats. And if you are shipping the same creative to more than one destination, the reserved-band rule is worth harmonising across all of them rather than maintaining a Pinterest-only template — the cross-platform labelling checklist is the five-minute version of that exercise.

FAQ

Does the AI modified label appear in the home feed grid?

Pinterest's documented placement is the bottom-left of the close-up view. Its help page does not describe a grid placement. Platform chrome changes without announcement, though, so treat that as the current documented behaviour rather than a permanent guarantee, and check your own Pins in both views when you ship a new template.

Can I avoid the label by stripping metadata from my export?

No. Metadata is only one of three triggers Pinterest describes; its classifiers are stated to work on content without obvious markers, and owner disclosure is a separate path. Stripping metadata removes the deliberate signal and leaves the automated one, which is the worst of both positions. Design for the label instead.

Does the label hurt distribution?

Pinterest publishes no data on the ranking treatment of labelled Pins, so anyone who tells you the number is guessing. What Pinterest does publish is a user-side control that lets people see fewer GenAI Pins in categories it describes as highly prone to AI modification — beauty, art, fashion and home decor. That is an audience-composition effect, not a ranking penalty, and it is measurable on your own account.

Should I write my own disclosure onto the Pin image?

On organic Pins, usually not — the platform AI content label is already doing that work and on-image text costs you design space. On ads, where the label is behind the ellipsis, a line in the description is cheap insurance. If you carry a formal obligation in your market, that decides it for you regardless of surface.