About the images on this site: AI-generated, and marked as such
Every photograph-like image on this site was produced by an image model, not a camera. What the marking means, how to recognise it, and the full list of frames.
- Frames
- 34 generated
- Photographs
- None
- Marking
- Page badge and metadata
- Logo and maps
- Drawn, not generated
Every large image on this site, the wide frame at the top of each page and the smaller frames on the place cards, was produced by an image model from a written description. None of them is a photograph. Each one carries a "AI-generated image" label placed over it by the page and a machine-readable provenance marker in the file metadata. The label is not burned into the pixels.
This page explains what that marking means, why the site uses generated images at all, and lists every frame so that nobody has to guess which is which. The short version: if it looks like a photograph of Dubai and it is on this site, it is a rendering, and the real place will differ from it in detail.
What the marking means
A frame that carries the "AI-generated image" label was created by a generative image model working from a text prompt. It is not a photograph of the place named in the caption, and it is not a survey or a reconstruction either. The model was never given plans, satellite imagery or reference photographs of the specific site. Where a frame resembles a real landmark, the Burj Al Arab on its island or the towers of Downtown Dubai across the water, the resemblance is the model's general idea of that landmark, and the proportions, the surroundings, the light and the small details are invented.
That matters for a practical guide more than it would for a magazine cover. A reader planning a visit may look at a frame and take from it where the beach is relative to the hotel, how wide a lane in the old quarter is, or how crowded a mall atrium gets. None of those readings is reliable from a generated image. The text on each page is written from published sources and is the part to plan from; the image sets a mood and nothing more.
Where the marks are, and why there are two
The visible mark is the "AI-generated image" label on the page, placed over the top edge of the frame. It is rendered by the site from the image record, so it remains visible when a frame is cropped for a phone or a wide screen and can be updated without altering the image itself.
The second mark is machine-readable. Each file carries a standard IPTC metadata field, DigitalSourceType, set to the value the standard defines for media produced by a generative model, trainedAlgorithmicMedia. Browsers do not show it, but tools built to check image provenance, and the systems behind image search, can read it without looking at the pixels.
The 18 wide frames
These are the frames at the top of pages. The same frame is used on more than one page where the subject fits, so the skyline frame appears on many of the practical pages, and the old-quarter lane on most of the culture and heritage pages. The description is what the frame shows, written from the frame and not from the page that happens to carry it.
- hero-beach: Empty sand and turquoise water, the Dubai Marina towers far off.
- hero-ain-dubai: Ain Dubai observation wheel on Bluewaters Island with the Marina skyline beyond.
- hero-atlantis-the-palm: Atlantis The Palm and its central arch viewed across the water.
- hero-burjalarab: The sail-shaped Burj Al Arab on its island, seen from the beach.
- hero-corridor: Glass towers on Sheikh Zayed Road, a metro viaduct crossing below.
- hero-creek: Wooden abras moored along Dubai Creek at dusk.
- hero-desert: A red dune ridge outside Dubai.
- hero-dubai-frame: The gold Dubai Frame rising above the lawns of Zabeel Park.
- hero-etihad-museum: The white Etihad Museum pavilion beside Union House and the UAE flag.
- hero-garden: A flower-covered arch over a path between dense petunia beds.
- hero-mall: A Dubai mall atrium under a vaulted glass roof.
- hero-mosque: A modern Gulf mosque: dome, one minaret, an arched colonnade.
- hero-museum-of-the-future: The Museum of the Future at blue hour, its steel ring lit through curved window openings.
- hero-old: A lamp-lit lane of wind towers in the old quarter.
- hero-palm-jumeirah: An elevated view of Palm Jumeirah showing the trunk, fronds and crescent.
- hero-park: A rollercoaster track above a theme park walkway at dusk.
- hero-pavilion: Lit market stalls under canopies, festoon lights above.
- hero-skyline: The Downtown Dubai skyline across water, the Burj Khalifa centre.
The 16 card frames
These are the smaller frames on place cards, in listings and in the "nearby" panels. Several of them are crops of a wide frame at a different aspect ratio; they are counted and marked separately because they are separate files.
- card-beach: Empty sand and turquoise water, the Dubai Marina towers far off.
- card-burj: The Burj Khalifa, tapering to its spire.
- card-burjalarab: The sail-shaped Burj Al Arab on its island, seen from the beach.
- card-creek: Wooden abras moored along Dubai Creek at dusk.
- card-downtown: The Downtown Dubai towers in haze, the Burj Khalifa centre.
- card-fahidi: A coral-stone lane in Al Fahidi, with a wind tower.
- card-garden: A flower-covered arch over a path between dense petunia beds.
- card-mall: A Dubai mall atrium under a vaulted glass roof.
- card-marina: Dubai towers at sunset above a long low-roofed building.
- card-mosque: A modern Gulf mosque: dome, one minaret, an arched colonnade.
- card-park: A rollercoaster track above a theme park walkway at dusk.
- card-pavilion: Lit market stalls under canopies, festoon lights above.
- card-souk: A carved wooden double door in a coral-stone wall.
- card-dubai-frame: The gold Dubai Frame rising above the lawns of Zabeel Park.
- card-etihad-museum: The white Etihad Museum pavilion beside Union House and the UAE flag.
- card-museum-of-the-future: The Museum of the Future at blue hour, its steel ring lit through curved window openings.
What is not generated
The site mark: an eye whose pupil is a tower silhouette over a waterline. Drawn by hand as inline SVG, not generated. Route maps and hour scales are rendered from data as SVG and HTML, not generated images. Neither carries a mark, because neither claims to show anything that exists: a drawn symbol and a chart are read as what they are. They are listed here so that the inventory on this page is complete.
There are no stock photographs, licensed photographs or reader-submitted photographs anywhere on the site. If that changes, this page will say so and the new material will be described separately, without a mark, because a mark on a real photograph would be an untrue statement about it.
Why generated images, and where to find real ones
The alternative was licensed photography, and it was considered. A practical reference to a city that changes as fast as Dubai does needs images that can be replaced quickly when a district is rebuilt or a landmark opens, and it needs them at a cost that a reference site can carry. Generated frames meet both conditions. Their weakness is the one this page exists to name: they look like photographs and are not. The marking is the price of using them honestly.
For photographs of the real places, the operators are the best source. Attraction pages on this site link to the operator or authority where one exists, and most of them publish current photography and, increasingly, live camera feeds. The Dubai Media Office publishes official imagery of the city, and the Dubai Department of Economy and Tourism maintains a large photographic library of attractions, districts and events. Where a detail in a generated frame and a photograph from one of those sources disagree, the photograph is right.
Questions about the images
Are any of the images on this site real photographs?
No. Every wide frame and every card frame was produced by an image model. The logo and the route maps are drawn graphics. There is no photography on the site at present.
Why does the label say "AI-generated image" on a picture of a real landmark?
Because the picture is a rendering of that landmark, not a photograph of it. The model produced its general idea of the building or district from a text description; proportions, surroundings and details are invented.
Can the images be reused elsewhere?
The frames carry an IPTC provenance field that should stay with the file. Anyone reusing one should add a visible disclosure and describe the image as generated, not as a photograph.
How can software tell that an image here is generated?
Each file carries the IPTC DigitalSourceType field set to trainedAlgorithmicMedia, the standard value for media produced by a generative model. Provenance tools and image-search systems read that field directly.