AI-generated images: what to check and disclose

Using AI-generated images on a business website? Here's what to check before you publish, when to tell readers, and how to stay honest about what an image shows.

Picdrome also offers original AI-generated images. This page is guidance for using AI-generated images from any source. The current license and terms at that source control what you may do with an image. This page is not legal advice.

Plain-language summary: An AI-generated image is made by an image-generation model, not captured with a camera. Before you publish one, check it for logos, readable brand text, trademarked products, and anything that looks like a real, identifiable person, and confirm the source's license covers your use. Use it honestly: don't pass off an AI-generated face as a real customer or team member, and check every image at full size before it goes live.

What an AI-generated image is

An AI-generated image is created by an image-generation model. It is not a photograph of a real event, place, or person. A scene that looks like a busy cafe, a product on a desk, or a person at a laptop was produced by a model, not taken with a camera.

That matters for how you use it. An AI-generated portrait is a synthetic likeness, not a depiction of an actual individual. An AI-generated "office" is an invented space, not a real company's premises. Treat an AI image as an illustration that fits a page, not as documentary evidence of something that happened.

What to screen for before you publish

Models can produce things you would never knowingly put on a business page. Before you use an AI-generated image, look for each of these.

Visible logos and brand text. Recognizable logos, readable brand names, and slogans can appear even when nobody asked for them. Don't put someone else's mark on your page.
Trademarked products. Identifiable trademarked products or distinctive packaging are a reason to choose a different image; generic, unbranded objects are safer.
Identifiable real people. Avoid images that appear to depict a specific, identifiable real person, including public figures.
Deceptive before/after. Avoid staged "before and after" imagery that implies a result your product or service can't honestly claim.
Political and sensitive content. Political messaging and sensitive or graphic depictions rarely belong on a general business page.
Readability and quality. Check that the image reads clearly at the size your page uses it and holds up as a background, hero, or section visual.

No check is perfect. Text-like shapes and logo-like marks are easy to miss at thumbnail size, which is why viewing the image at full size matters. See Limits and artifacts below.

Using AI images responsibly

Responsible use is mostly about being honest with your visitors.

Don't present an AI person as a real person. An AI-generated face is not a real customer, team member, founder, or reviewer. Don't attach it to a testimonial, a quote, a staff bio, or a headshot that implies a specific real individual stands behind it.
Keep claims truthful. Pairing an invented image with a factual statement is fine; using it to imply something untrue (a result, an endorsement, a relationship) is not.
Consider disclosing AI imagery where it matters. In editorial, news, or contexts where the realness of an image carries weight, it's worth telling readers the visual is AI-generated. For a decorative hero or a section background, disclosure usually isn't expected.
Follow the rules for your context. Some platforms, marketplaces, and regions have their own labeling expectations for synthetic media. Check what applies to you.
Check the source's terms. The license or terms where the image came from decide what you may do with it, including commercial use.

Not sure whether to label an image as AI-generated? Our guide Should you disclose AI-generated images? walks through when disclosure helps and how to do it cleanly. For getting a natural look without misleading anyone, see making AI images look natural.

Limits and artifacts

AI image generation is impressive but imperfect. Images can contain artifacts: hands or fingers that look off, garbled text, warped objects, odd reflections, or background details that don't quite make sense. These can be subtle and easy to miss until an image is published at full size.

Before you publish:

View the image at the size you'll actually use it.
Check faces, hands, and any text-like areas for distortion.
Confirm the image fits the message of your page and won't mislead.
Read the current license or terms at the image's source so you know what your use covers.

A quick check takes seconds and saves you from shipping an image with an obvious glitch. You publish it, so you get the final say on whether it's good enough.

Who this is for

This page is written for the people most affected by AI imagery on business sites: solo founders and small-business owners placing images on their own sites, freelancers and agencies sourcing visuals for client work, and creators and marketers using images across landing pages, blogs, and emails. If you publish images for others or in a context where realness matters, the responsible-use and disclosure notes above are the ones to read closely.

About this page: Picdrome is published by Common Syntax LLC. It publishes practical guidance and original AI-generated images; for the standards behind the guidance, see the editorial policy.

Frequently asked questions

Are AI-generated images real photographs?

No. They are made by image-generation models, not captured as photographs of real events, places, or people, even when they look photorealistic. Treat them as illustrations, not documentation.

Could an AI image accidentally show a real person or logo?

Yes. Models can produce logo-like marks or face-like results unintentionally, which is why you should check every image at full size before publishing. If something looks like a real mark or a real person, choose a different image.

Do I have to tell visitors my images are AI-generated?

Often you don't, especially for decorative heroes and backgrounds. It matters more in editorial, news, or contexts where the realness of an image carries weight, and where a platform or region asks for labeling. When in doubt, see Should you disclose AI-generated images?

Are AI images safe to use commercially?

That depends on the source's license or terms, not on the fact that the image is AI-generated. Read them before you publish, and see using AI images commercially for what to check. "Safe" also depends on how you use the image: keep your claims truthful and don't present an AI-generated person as a real customer or team member. This isn't legal advice.

What if I spot an artifact or a problem in an image?

Don't publish an image you're not happy with. Choose another, or fix the problem before it goes live.

Read the details

Want images that read as natural without misleading anyone? See how to make AI images look natural.