"Should I tell people my website images are AI-generated?" sounds like one question with a yes-or-no answer. It's actually three questions stacked on top of each other, and they don't all point the same way. This walks through each layer so you can decide your own stance instead of acting on a rule that may not exist.
A quick frame before the steps: this is general guidance, not legal advice. AI-labeling expectations are changing, and what applies to you depends on where you operate, what kind of site you run, and how you use the images. For anything with real legal or financial stakes, talk to a professional. For how Picdrome handles its own images, the AI image disclosure page is the worked example throughout.
The three questions hiding inside "should I disclose?"
When someone asks whether they need to disclose AI imagery, they're usually mixing three separate things:
- Am I legally required to? A question about law and regulation in your jurisdiction.
- Does a platform or ad network require it? A question about the terms of services you use to publish or advertise.
- Is it the right thing to do for my visitors? A question about trust and ethics, where you get to make a judgment call.
These can have different answers. You might face no legal requirement, no platform requirement, and still choose to add a small note because it fits how you want to treat your audience. Or a platform might ask for a label even where the law says nothing. Collapsing all three into a single "you must" or "you don't have to" is where people get it wrong.
The goal here is to keep the layers separate so your decision is deliberate.
Layer 1: Legal and regulatory rules
This is the layer most likely to change between when this is written and when you read it.
Rules around labeling synthetic or AI-generated media are being written and revised in various places, and they tend to be specific: they may apply to certain contexts (political ads, deepfakes of real people, news media) more than to a decorative hero image on a small-business homepage. Some regions are more active than others. What counts as "AI-generated" and what triggers a labeling duty is not settled and varies by jurisdiction and context.
A few honest takeaways for a small site:
- There may be no general law requiring you to label an ordinary decorative image as AI-generated, and there may be specific rules that bite in narrower situations. Both can be true at once.
- Rules aimed at synthetic media usually care most about images that could mislead about a real person or a real event. That maps directly onto the one rule that never moves (see below).
- Because this is a moving target, treat any confident claim you read online about "the law on AI images" with caution, including this one. Check current sources for your location, and for real stakes, get advice from a professional.
If you operate in a regulated industry, run political or issue advertising, or publish anything that resembles news or documentary content, this layer deserves more attention than a typical brochure site needs.
Layer 2: Platform and ad-policy requirements
This layer is separate from the law and often more concrete in day-to-day terms.
Many platforms — social networks, ad networks, app stores, some marketplaces — have their own rules about labeling AI-generated or synthetic content, and some provide a toggle or field for it. These rules live in each platform's policies, change on their own schedule, and apply because you agreed to the terms, not because a court said so. Exact requirements differ by platform and by content type, and they get updated.
Practical way to handle it:
- If you're running ads or publishing on a third-party platform, read that platform's current content and advertising policies for AI or synthetic-media labeling. Each one is its own answer.
- If you're publishing on your own website, no third-party policy forces your hand. The decision moves to Layer 3.
- Don't assume one platform's rule applies everywhere. A label requirement on a social feed says nothing about your homepage.
The short version: platform rules can require disclosure in places the law doesn't, and they're easy to check because they're written down in the policy you accepted.
Layer 3: Trust and ethics, as a choice
Where there's no legal or platform requirement, disclosure becomes a choice about how you want to treat the people reading your site. This is the layer you fully control, and there's no universal right answer.
A useful test is whether the realness of the image carries weight in context:
- A decorative hero, a section background, or an abstract texture isn't pretending to document anything. Most visitors don't assume it's a photograph of your actual office, and disclosure usually isn't expected.
- An image that stands in for something real — a photo-style shot implying it's your team, your premises, your product, or a customer — is where honesty matters more, because a viewer might reasonably read it as documentary.
There's one place this stops being a judgment call and becomes a firm rule, on Picdrome and in general:
Never present an AI-generated person as a real customer, team member, founder, or reviewer. Don't attach an AI face to a testimonial, a staff bio, a headshot, or anything that implies a specific real individual stands behind it.
That isn't about labeling preference. It's about not misleading people, and it holds regardless of what any law or platform says. You can keep an AI hero image undisclosed and still be completely honest, as long as you're not using it to claim something untrue.
Practical disclosure patterns for a small site
If you decide some disclosure fits, you don't need a banner or a pop-up. Match the visibility of the disclosure to how much the image's realness matters. Here are patterns that work on an ordinary business site, from lightest to most explicit.
| Pattern | Where it lives | Best for | Effort |
|---|---|---|---|
| Image caption | Directly under or beside one image | A specific image where realness matters (an "about" visual, a case-study illustration) | Low |
| Image-credits line | A short credit near the image or in a sidebar | Blog posts and articles where you already credit sources | Low |
| Footer AI-use note | One sentence in the site footer | Whole-site coverage without touching every page | Low |
| Dedicated disclosure page | A standalone page linked from the footer | Sites that want a clear, reusable explanation in one place | Medium |
Wording can be plain and short. A caption might read: "Illustration generated with AI." A credits line: "Images: AI-generated, used for illustration." A footer note: "Some images on this site are AI-generated and used for illustration, not as photographs of real people or events."
Two things keep any pattern clean:
- Say it once, where it's findable. A site-wide footer note plus a caption on the few images that matter is usually enough. You don't need to label every background texture.
- Don't let the disclosure imply more than it should. "AI-generated illustration" is honest. Avoid wording that tries to pass an invented scene off as a real one.
For the deeper rights questions that sit alongside disclosure, the pillar guide on using free and AI-generated images safely ties this together, and can you use AI-generated images commercially? covers the commercial-use angle specifically.
How Picdrome discloses its own AI images
Picdrome publishes an AI-generated image library, so it has to answer this question for itself. It's offered here as one transparent example of the layers above, not as a template you're obliged to copy.
What Picdrome does:
- States the provenance plainly. The AI image disclosure page says, without hedging, that every image is AI-generated rather than photographed, and explains what gets screened out (visible logos, readable brand text, trademarked products, identifiable real people without clear rights).
- Holds the non-negotiable line. It tells users not to present an AI-generated person as a real customer or team member — the one rule that applies regardless of jurisdiction.
- Separates "can use" from "how to use honestly." The images are free to use for personal and commercial website projects, subject to the Picdrome Image License, and the disclosure page focuses on using them truthfully.
- Writes down its standards. The editorial policy sets out what Picdrome publishes and why, so the disclosure isn't a one-line afterthought.
If you want a model for your own approach, that combination — a plain statement of provenance, a firm rule against faking real people, and a clear link to your terms — is a reasonable shape to borrow. When trademark and copyright questions come up alongside disclosure, how to avoid copyright and trademark trouble covers what to screen for in the image itself.
See Picdrome's own approach on the AI image disclosure page and decide what fits your site.
In Plain English
"Should I disclose AI images?" is three questions, not one: a legal/regulatory question (evolving, varies by place and context, and worth checking current sources), a platform/ad-policy question (read the terms of whatever you publish or advertise on), and a trust-and-ethics question (your call). This helps anyone running a small business site with AI imagery who's unsure whether a note is required or just good practice. Use a small disclosure — a caption, an image-credits line, or a footer note — when an image's realness carries weight, and lighten up on purely decorative visuals. The rule that never changes: never pass off an AI-generated person as a real customer or team member. This is general guidance, not legal advice; for specifics, check your jurisdiction, the relevant platform policies, and the Picdrome Image License, and get professional advice when the stakes are real. Next step: look at how the AI image disclosure page handles it, then pick the lightest pattern that's honest for your site.
Common Questions
Do I legally have to label AI images on my website?
It depends on where you operate and what the image does. There may be no general requirement for ordinary decorative images, while specific rules can apply to narrower contexts like political ads or media that could mislead about a real person. The rules are evolving and vary by jurisdiction, so check current sources for your location and get professional advice for real stakes.
Where should a disclosure go on a small site?
Match it to how much the image's realness matters. A site-wide footer note covers everything lightly; a caption or image-credits line handles the specific images where realness carries weight, such as an "about" page visual. You rarely need more than one footer note plus a few captions.
Is it dishonest to use AI images without labeling them?
Not by itself. A decorative hero or background that isn't pretending to document anything can be undisclosed and still honest. It becomes a problem only when an image is used to imply something untrue — most clearly, presenting an AI-generated person as a real customer or team member, which you should never do.