Aipedia: Landscape architects

AI visualisations and LI TGN 06/19

Aipedia, by the Apex broking team · Last reviewed 2026-10-30

AI images can only count as technical visualisations under TGN 06/19 if the process is replicable and the proposal accurately sized and positioned, and appeal use must be declared. Visuals are often the most contested part of landscape evidence.

Part of: AI in Landscape architects

In short

TGN 06/19 asks for "a fair representation of what would be likely to be seen", with four visualisation types from annotated photographs to survey-verifiable photomontages. Types 3 and 4 need a technical methodology and camera data, compositing must be done with "visual skill, judgement and objectivity", and the proposal's size and position must be accurate. Generative AI struggles to meet those tests. The Planning Inspectorate requires generated or altered images to be declared. A misleading visual can lead to challenge at inquiry, a costs application and a PI claim, subject to the wording.

How landscape practices use AI for visuals

AI image tools are used across the visual workflow. Practices generate early concept images from sketches or text prompts, use AI to mask foregrounds, replace skies or add people, and simulate how planting will look at year one and year fifteen. Some tools can offer one-click enhancement of a rendered model view. The LI's Campus session on AI visualisation in December 2024 covered image generation, generative design and presentation graphics. The difficulty is that the same tools used for a mood board can end up in a planning submission.

What TGN 06/19 requires

LI Technical Guidance Note 06/19 (September 2019, under review since January 2024) does not mention AI, but its tests apply to any visual used to assess or present a development proposal:

visualisations should give "a fair representation of what would be likely to be seen" (1.2.9);
four types: Type 1 annotated viewpoint photograph, Type 2 3D wireline or model, Type 3 photomontage or photowire, Type 4 survey or scale-verifiable photomontage (3.5.2, Table 2);
illustrative images such as marketing perspectives and bird's-eye views are not technical visualisations (1.2.3);
Types 3 and 4 need a Technical Methodology and per-viewpoint camera data (4.4.8, 4.5.6, Appendix 10), with the type labelled on each image;
compositing must be "carried out with visual skill, judgement and objectivity" (6.1.1) and the size and position of the proposal must be accurate (6.1.2);
processes should be transparent and replicable (2.3), and the choice of type proportionate to sensitivity (3.5.10).

Can an AI image meet the standard?

A purely generative image cannot meet the Type 3 or 4 tests, because there is no camera data or methodology that would let a third party replicate it and no guarantee the proposal is the right size in the right place. It may be acceptable as an illustrative image if it is clearly labelled as such and not used to judge visual effects.

AI used as a tool inside a conventional verified workflow is a different case. Automated masking or tree placement within a model-based photomontage can be compatible with TGN 06/19 if the practitioner checks the result, records the step in the methodology and the output remains accurate. The risk is AI enhancement that quietly softens a scheme: fuller planting, hazier distance, a brighter sky or a smaller building edge.

The Planning Inspectorate declaration

The Planning Inspectorate's guidance on AI in casework evidence (updated 23 September 2026) requires parties to declare intentional AI use to generate or alter images or videos, identify which parts, explain what checks were made and confirm responsibility for accuracy. It discourages AI cartoons or exaggerated illustrations of a proposal, and improper use may be unreasonable behaviour leading to a costs award. Each expert declares for their own documents, so the landscape witness answers for their own visuals. The planning consultant's side is covered in AI visualisations and photomontages.

What can go wrong

Illustrative scenario. A practice prepares Type 3 photomontages for a housing appeal. To save time, a technician uses an AI tool to "tidy" the composited views, and it extends existing hedgerows and thins the roofline. The images are submitted without an AI declaration. At inquiry the opposing landscape witness overlays a wireline from the same camera position and shows the development is larger than depicted. The inspector gives the visuals little weight, the appeal fails and the appellant seeks the costs of the wasted appeal from the practice.

The same error in an LVIA chapter could undermine the whole assessment, and in marketing imagery it raises truthful advertising issues.

How PI usually responds

A client's claim that negligent visualisations caused a lost appeal or wasted costs is a civil liability claim arising from professional services, and landscape PI usually responds, subject to the wording. Insurers may look at whether the practice followed TGN 06/19 and its own methodology, whether AI use was disclosed at renewal, and whether the loss was really caused by the visuals or by the merits of the scheme. A costs award made against the appellant is the client's liability; a later claim by the client against the practice is what PI addresses. See does landscape architects' PI cover AI errors? and PI for landscape architects in 2026.

Practical controls

Practices can keep AI out of technical visuals, or keep it under control, with a few steps:

label every image with its TGN 06/19 type, and mark AI-generated concepts as illustrative;
record any AI step in the Technical Methodology, with the tool's function and the check made;
overlay a wireline on any AI-assisted photomontage before issue to confirm size and position;
ban generative fill on Type 3 and 4 images unless a senior practitioner signs it off;
prepare the Planning Inspectorate AI declaration alongside the visuals, not afterwards;
keep source photographs, camera data and unedited composites on file.

How Apex helps

Apex arranges PI for landscape practices that produce visual evidence. A broker can check how your wording treats claims over visualisations and help you describe your AI workflow to insurers. See the AI in landscape architecture overview and the LI Code and AI guide.

Sources

Frequently asked

Can I use AI images in a planning application?

You can use AI images as illustrative material if they are clearly labelled and not presented as technical visualisations. For appeals and NSIP examinations, any intentional AI use to generate or alter images must be declared under Planning Inspectorate guidance, saying which parts and what checks were made.

Does TGN 06/19 mention AI?

No. TGN 06/19 was published in September 2019 and has been under review since January 2024. Its tests still apply: a fair representation, a declared visualisation type, a technical methodology and camera data for Types 3 and 4, and accurate size and position of the proposal.

Is AI masking in a photomontage allowed?

TGN 06/19 does not rule out digital tools. Compositing must be done with visual skill, judgement and objectivity, and processes should be transparent and replicable. If an AI masking step is checked, recorded in the methodology and leaves the proposal accurately sized and positioned, it can be compatible with the guidance.

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.