Aipedia: Architects

AI renders and visualisations: the risks for architects

Aipedia, by the Apex broking team · Last reviewed 2026-10-27

AI renders are lawful to use, but an architect must not let them mislead clients, planners or the public about what will be built. Visualisation is the most common AI task in practice, so it is where expectation disputes start.

Part of: AI in Architects

In short

About 56% of AI-using practices use it for early-stage design visualisation (RIBA AI Report 2026). Architects Code Standard 1.2 requires accurate promotion and RIBA's Code bans misleading statements "through words or images". The Planning Inspectorate requires AI-generated or altered images in appeal evidence to be declared, with the parts identified. If a client says they relied on an image that the design, budget or consent could never deliver, the claim usually lands on PI, subject to the wording.

How architects use AI images

Generative image tools turn a sketch, massing model or text prompt into a convincing render in minutes. Practices use them for competition entries, client workshops, early option studies, marketing and planning material. The RIBA AI Report 2026 found early-stage design visualisation is the leading AI task, used by about 56% of practices that use AI.

The speed is the risk. A generative image fills gaps with plausible detail: materials, glazing ratios, planting, light and context that nobody designed or costed. A client who falls for the image may treat it as a promise.

The duties that apply

Architects Code Standard 1.2: promote services responsibly and accurately. Showing AI images of work the practice did not design, or images the design cannot achieve, risks breaching it.
Standard 5.3: explain your role and manage expectations, including what an early image does and does not represent.
RIBA Code, Principle 3, 9.2(a): no misleading statements "through words or images" (see the RIBA Code of Professional Conduct).
Planning Inspectorate: its guidance on AI in casework evidence (updated 23 September 2026) requires parties to declare intentional AI use to generate or alter images or videos, saying which parts, what checks were made and confirming responsibility for accuracy. It discourages AI cartoons or exaggerated illustrations, and improper AI use may be unreasonable behaviour leading to a costs award.

The ARB and RIBA guide covers the wider Code, and AI design and access statements covers the written side of planning submissions. For the wider picture, see the AI in architecture overview.

What can go wrong

Illustrative scenario. A practice wins a residential commission using AI renders showing large frameless glazing, natural stone and a mature landscape. The fee proposal and cost plan assume standard windows and render. During detailed design the client learns the glazing would fail overheating checks and the stone is unaffordable. The client says they appointed the practice and bought the site on the strength of what they were shown, and claims the wasted costs.

Other routes to a claim include AI-altered verified views that understate the scale of a scheme, an image used in an appeal without a declaration, and AI-generated context that shows neighbouring land as open space. Many negligence claims against architects turn on what the client was told; images are now part of that.

How PI usually responds

A claim that a client relied on misleading images in the course of professional services will usually be treated as a civil liability claim under an architects' PI policy, subject to the terms. Points that commonly arise:

whether the image was part of professional services or pure marketing, which some wordings treat differently;
whether the loss flows from the image or from the client's own decisions, which goes to causation and scope of duty;
a Planning Inspectorate costs award against the client may be pursued against the architect, and whether PI responds depends on the wording.

ARB requires at least £250,000 each and every claim under its PII guidance. Copyright in images raises separate issues, covered in generative design and copyright.

Practical controls

Label AI-generated images as illustrative on the image itself, not only in a footnote.
Do not use AI images of buildings the practice did not design in bids or marketing (see AI marketing for practices).
Check each image against the brief, budget and known constraints before a client sees it.
Keep the prompt, tool and date with the file, and record who approved it.
Declare AI-generated or altered images in appeal evidence, image by image.
Use verified, non-AI methods for views that will be relied on to assess impact.

How Apex helps

Apex arranges PI for architects. A broker can check how your wording treats claims arising from visualisations and marketing material, and whether planning costs exposures are within cover.

Sources

Frequently asked

Must I tell clients a render is AI-generated?

No rule specifically requires it, but Standard 1.2 requires accurate promotion and Standard 5.3 managing expectations. Labelling AI images as illustrative and explaining what they do not show is the simplest way to meet both and reduces the risk of a reliance claim.

Do I have to declare AI images at a planning appeal?

Yes, where the Planning Inspectorate guidance applies. Intentional AI use to generate or alter images must be declared, identifying which images, what checks were made and confirming responsibility for accuracy. Each expert should declare for their own documents.

Is a misleading render a PI claim?

Usually it can be, where the client alleges negligent professional services and a financial loss. Whether the policy responds depends on the wording, whether the image was professional work or marketing, and whether the loss was caused by reliance on it.

Check your PI covers how you use AI

Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.

Start a PI quote Request a callback

Apex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.