Direct answer: yes, but only to present the real place well — never to show a place, room, view or facility that guests will not get. UK advertising rules apply to AI imagery exactly as they apply to photography and Photoshop: the CAP Code “is media‑neutral” and the rules apply “regardless of how the content was created” (ASA, 11 June 2026). The ASA’s travel guidance says marketers “should also ensure that the photos which they feature in their ads are representative” (ASA travel marketing guidance, 28 January 2026). Under the Digital Markets, Competition and Consumers Act 2024 (DMCC Act), an “overall presentation which is likely to deceive the average consumer” is a misleading action even if the information in it is true (DMCC Act, section 226). And the platforms you sell through are explicit: Airbnb will ask hosts to remove content where AI has been used to “edit flaws, hide damage, add amenities or attributes that are not part of a listing” (Airbnb ground rules), and Expedia Group bans “AI-generated or AI-edited content to misrepresent a property” (Expedia Group partner provided content policy).
Kaize is a UK AI product studio for travel and hospitality. This article is based on ASA and CAP guidance and rulings, the DMCC Act and CMA guidance, the EU AI Act text and Commission material, and the photo and content policies of Airbnb, Expedia Group, Booking.com and Google, read on 15 October 2026. It is not legal advice or a case study; the controls below are our recommendations.
Why this question is live in October 2026
Generative image tools are now built into the software hotels, villa companies and tour operators already use: photo editors that remove objects or replace skies, “virtual staging” that furnishes an empty room, tools that expand a cropped shot into a wide one, and generators that produce a lifestyle scene or a destination image in seconds. The pressure to use them is commercial — new listings, refurbishments, seasonal campaigns and dozens of OTA galleries to fill — and the outputs are often indistinguishable from photographs.
Regulators have not created AI-specific image rules in the UK. CAP said in May 2025 that “there is no blanket legal requirement in the UK to disclose the use of AI in ads”, but that existing rules apply however content is generated, and that “disclosure alone is very unlikely to mitigate the harm caused by a fundamentally misleading message” (CAP, 29 May 2025). In June 2026 it added that when an AI tool produces something problematic, “the advertiser remains responsible for any output used in their ads” — including where ads are produced with a platform’s automated tools (ASA, 11 June 2026). Meanwhile, the EU AI Act’s transparency rules in Article 50, including disclosure of deep fakes, started to apply on 2 August 2026 (AI Act Service Desk timeline).
The test UK regulators actually apply to travel photos
The ASA does not ask whether a photo is flattering. It asks whether it gives a fair impression of what the consumer will receive. The clearest travel example is an old ruling on a Maldives resort: a guest complained that the website photos hid sandbags on the beach. The ASA did not uphold the complaint because consumers would understand that photographs “would reflect the hotel and resort in the best possible light”, and the images “gave a fair impression of the general appearance of the resort and the beaches” (ASA ruling, Hayes & Jarvis (Travel) Ltd, 4 March 2015). Best light is allowed. A different place is not.
Three further principles from ASA guidance matter for AI imagery. Images that give an inaccurate impression of what consumers will receive — by “featuring the wrong product, including extras, or by exaggerating the quality or size of the product” — are likely to mislead; and if using pre- or post-production techniques, images “should not be altered in a way which exaggerates what the product can do” (ASA, 11 September 2025). Descriptions of accommodation, facilities and amenities must not mislead — the ASA upheld a complaint about a hotel described as a “Secluded Haven” during construction works (ASA travel marketing guidance). And marketers must hold evidence for objective claims before publishing (CAP Code rule 3.7) (ASA travel marketing guidance). A photo of a sea view is an objective claim that the room has one.
The DMCC Act provides the legal backbone. A commercial practice is unfair if a misleading action is likely to cause the average consumer to take a transactional decision they would not have taken otherwise (DMCC Act, section 225), and a misleading action includes false or misleading information and a deceptive overall presentation (DMCC Act, section 226). The CMA’s guidance explains that the overall presentation test catches information that “is not actually false, but is arranged, displayed or otherwise presented in a way that is likely to deceive” (CMA207 unfair commercial practices guidance). Under the new regime, if a company infringes consumer protection law, the CMA can fine it up to 10% of global turnover (CMA, 27 March 2026). For a travel booking, the transactional decision is usually the booking itself, and the photo is often what drove it.
Five uses of AI imagery, from low to high risk
1. Technical correction of a real photo
Straightening verticals, correcting exposure, white balance and noise, or sharpening a real photo of the real room is ordinary post-production. CAP’s own examples of AI making “subtle non-material changes” include “removing unwanted background elements” and “improving the weather conditions” (CAP, 29 May 2025). Our view: fine, with one caution for travel. On a beach resort, a ski chalet or a rooftop bar, weather and setting can be material; a grey sky turned blue is cosmetic, a rough sea made calm on a page selling a swimming beach may not be. Platforms also push back on heavy processing: Booking.com asks partners to avoid “over-processed photos” and “photos with filters, strong contrast, over-saturations or excessive HDR” (Booking.com photo requirements), and Google Business Profile says photos should have “no significant alterations or excessive use of filters or AI. In other words, the image should represent reality” (Google Business Profile photo guidance).
2. Removing things that are temporarily there
Erasing a cleaning trolley, a parked van or a guest’s suitcase from a real photo is low risk. Erasing things that are permanently or seasonally there — scaffolding, a building site next door, a motorway, power lines over the pool, a neighbouring nightclub, beach erosion defences that are visible most of the season — is where AI editing becomes a misleading presentation. The question for each removal is simple: would a guest arriving in the booked dates see it? If yes, it stays in the picture, and the ASA “Secluded Haven” example shows that disruption can be material even when the photos are real.
3. Virtual staging, extended views and “improved” rooms
Furnishing an empty room, widening a frame beyond what the lens captured, adding a terrace view, upgrading the bathroom finishes, or making a standard room look like the suite next door all change the product. These are the cases the platforms name. Airbnb requires photos to “accurately represent the space provided” and will ask hosts to remove AI content that edits flaws, hides damage or adds amenities (Airbnb ground rules). Expedia Group prohibits content that “exaggerates or omits material details that would affect a traveler’s expectations” and requires partners to update content promptly “when material features, conditions, or amenities change” (Expedia Group partner provided content policy). Booking.com does not allow images that “don’t relate to the actual travel experience” or “AI-generated content that is misleading or deceptive” (Booking.com content guidelines). Our view: do not use virtual staging or AI-extended frames for any room type, view or facility a guest can book. If the room is not ready, photograph it when it is.
4. Fully synthetic images of a real property
An AI-generated “photo” of your own hotel, villa or tour vehicle is the highest-risk use. It will look like a photograph; the average consumer will treat it as one; and it shows something that was never captured. Under the EU AI Act, AI-generated or manipulated image content that “resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful” is a “deep fake” (AI Act, Article 3(60)), and deployers who generate one must disclose that the content has been artificially generated or manipulated (AI Act, Article 50(4)). That obligation can reach UK businesses: the Act applies to deployers in third countries “where the output produced by the AI system is used in the Union” (AI Act, Article 2). Take advice on scope if you market to EU consumers. Under UK rules a label does not fix the problem: CAP says it would “almost certainly be against the rules” to make a misleading claim and then rely on disclosing AI use to disclaim it (CAP, 29 May 2025).
There is one legitimate version: pre-opening or pre-refurbishment visualisations. Architects’ renders have long been used to sell properties that do not exist yet. Our recommendation is to label them on the image itself as a computer-generated visualisation, keep them off OTA galleries and Google Business Profile, and replace them with photographs before the first guest arrives.
5. Synthetic mood, destination and lifestyle imagery
A generated image of a generic couple on a generic terrace, used on a brand campaign that does not depict a specific property, is the lowest-risk synthetic use — but it is not risk free. If it sits on a page for a specific hotel, villa or tour, consumers may read it as that hotel, villa or tour. Destination images carry the same trap: a crowd-free, generated version of a famous viewpoint on a tour page implies an experience the tour may not deliver. CAP also warns that AI tools can produce “harmful, discriminatory stereotypes”, so generated people need the same review as any casting decision (ASA, 11 June 2026). And a generated person must never appear as a guest giving a testimonial; that is a fake review problem, covered in our article on AI and guest reviews under the DMCC Act.
Where platform rules are stricter than the law
UK law asks whether the overall impression misleads. The platforms that supply most hotel and villa bookings ask for more, and their rules decide whether your listing stays live:
- Airbnb: photos and descriptions must accurately represent the space; AI or digital edits that hide flaws or damage, add amenities or otherwise misrepresent the listing must be removed; repeated or severe violations can lead to suspension, loss of Superhost status or payout adjustments (Airbnb ground rules).
- Expedia Group: the partner content policy covers “any AI-generated content”, bans AI-generated or AI-edited content that misrepresents a property, and requires disclosure of significant disruptions such as ongoing construction or maintenance (Expedia Group partner provided content policy).
- Booking.com: photos should be property-related and “set accurate expectations for your guests”, without over-processing, heavy filters or distortion (Booking.com photo requirements); misleading or deceptive AI-generated content is not allowed (Booking.com content guidelines).
- Google Business Profile: photos should have no significant alterations or excessive use of filters or AI, and “should represent reality” (Google Business Profile photo guidance). In Google Merchant Center, Google already requires that images created with generative AI carry the IPTC DigitalSourceType TrainedAlgorithmicMedia metadata tag and that this metadata is not removed (Google Merchant Center help) — a sign of where labelling is heading.
Kaize reading: design to the strictest channel you depend on. A gallery that is arguably lawful but gets a listing flagged on Airbnb or Expedia is still a commercial failure, and one master image library usually feeds every channel.
Provenance: the control most image pipelines are missing
Most operators cannot answer a simple question: which of our published images were edited or generated by AI, and how? That matters for complaints, for platform disputes and, for EU-facing marketing, for Article 50. It is also increasingly machine-readable. The IPTC recommends that software creating images with trained AI models writes the Digital Source Type value “trainedAlgorithmicMedia” into the image metadata, or into a C2PA manifest (IPTC). Under the AI Act, providers of generative AI systems must mark outputs “in a machine-readable format and detectable as artificially generated or manipulated” (AI Act, Article 50(2)); for certain providers whose systems were already on the market before 2 August 2026, the transitional deadline is 2 December 2026 (AI Act Service Desk timeline). The Commission published a voluntary Code of Practice on marking and labelling AI-generated content on 10 June 2026 (European Commission, 10 June 2026).
The practical consequence for a hotel or tour operator is that editing tools, image CDNs and booking engines can strip this metadata on export, resize or upload. If your process removes the evidence of how an image was made, you cannot show a platform, the ASA or a guest what you did. Keep originals, keep edit logs, and test whether your pipeline preserves metadata.
A worked example: a 30-villa group refreshing its galleries
This example is hypothetical. A villa company wants new galleries for 30 properties before peak season. Its photographer covers 22 villas; for the other 8, the marketing team proposes AI: sky replacement and object removal on older photos, virtual staging for two newly refurbished villas still awaiting furniture, an AI-extended pool shot for a villa with a small terrace, and a generated sunset image for each villa’s hero slot.
Applying the tests above: sky replacement and removing a hose pipe are fine if the views and surroundings are unchanged. Virtual staging is out for bookable villas — re-shoot once furniture arrives, or publish without those rooms and say so. The extended pool shot is out: it exaggerates size, the exact case ASA guidance warns about. The generated sunset hero is out on each villa’s own page because it will be read as that villa; it can stay on a brand page about the destination if it does not depict a specific property. One villa has a building site next door until spring: that goes in the description and on every channel that asks for disruption disclosure, whatever the photos show. Every AI-touched image is logged with the tool, the edit and the approver before it goes to any channel.
A pre-publish gate for AI imagery (Kaize recommendation)
This is our recommended gate before any AI-edited or AI-generated image is published for a property, room, tour or experience. It is not an ASA or CMA standard or legal advice; agree it with your marketing, legal and distribution owners.
- A material-features list for each property and product — room size and layout, view, bathroom, pool, beach access, terrace, accessibility features, surroundings — that no AI edit may add, enlarge, remove or improve.
- Allowed edits written down: exposure, colour, straightening, noise, and removal of temporary objects. Everything else needs a named approver.
- No virtual staging, AI-extended frames or fully synthetic images for anything a guest can book. Pre-opening visualisations labelled on the image and kept off OTA and Google galleries.
- Known disruption — construction, refurbishment, closed facilities, seasonal works — disclosed in text on every channel, regardless of what the photos show.
- An image register: original file, tool used, edit description, approver, date, channels published to. Provenance metadata preserved on export and tested through your CDN and channel manager.
- EU-facing marketing reviewed for AI Act Article 50(4): any AI-generated or manipulated image that could appear to be an authentic photo of a real place is either replaced or disclosed.
- Generated people reviewed for stereotypes and never presented as real guests or reviewers.
- Re-shoot or review triggers: refurbishment, change of operator, new neighbouring development, or a guest complaint that the photos do not match.
What this article is not
- Not legal advice, and not a review or ranking of any photo-editing or image-generation product.
- Not a claim that AI image tools are banned for travel marketing — UK rules apply to outcomes, not tools.
- Not a finding that any business named in an ASA ruling broke the rules; the 2015 Maldives complaint was not upheld.
- Not a case study; Kaize reports no client results here.
Questions
Do we have to label every AI-edited photo in the UK?
No. CAP says there is no blanket UK requirement to disclose AI use in ads; the question is whether consumers would be misled without disclosure, and whether a disclosure clarifies the ad or contradicts it (CAP, 29 May 2025). Platform and EU rules can go further.
Has the ASA ruled on an AI-generated hotel photo?
We found no ASA ruling on an AI-generated hotel or holiday image in the guidance and rulings we read. In 2023 the ASA said it had “yet to rule, to their knowledge, on any ads using AI generated images” (ASA, 31 August 2023); by 2026 it was citing rulings on AI-generated images in other sectors (ASA, 11 June 2026). The existing travel photo principles apply either way.
Can we use an AI image of the destination rather than the hotel?
Often, yes, on brand and inspiration pages, as long as it does not depict a specific property, tour or experience you sell, and does not imply conditions — empty beaches, clear views, wildlife — that the product cannot reasonably deliver.
Kaize point of view
AI may polish a real photo; it must not invent the product. Image tools are useful in travel marketing — faster editing, consistent galleries across channels, cleaner shots of real rooms — and none of that requires a model to add a view, remove a building site or furnish a room that is still empty. Our view is that hotels, villa companies, tour operators and DMCs should treat imagery as product data, not decoration: a defined list of material features AI cannot touch, a register of every AI-touched image, provenance kept intact, and a person accountable for each one. Do that and AI makes galleries better and cheaper; skip it and the photos become the complaint.
Related reading: AI and guest reviews under the DMCC Act, AI green claims in UK travel marketing, AI disclosure in guest chat under Article 50, short-term let registration before AI publishes listings, accessibility of AI booking journeys and AI agents and UK consumer law.
If you run a hotel group, tour operator or DMC and want help designing an image pipeline, channel content rules or an AI editing workflow that stays within ASA, DMCC and platform rules, start with an AI opportunity review or discuss a project with Kaize. For how content flows between your PMS, channel manager and booking system, see connecting travel booking systems and CRM and what stops hotel and villa overbooking.