Aipedia: Landscape architects

AI tree surveys and arboricultural constraints

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

AI can map trees and canopies from drone, LiDAR and aerial data, but it cannot reliably identify species, condition or protected status, and root protection areas built on it need checking. Tree constraints shape layouts, so errors spread through a scheme.

Part of: AI in Landscape architects

In short

Tree detection from point clouds and imagery may be used at feasibility stage. AI models can misplace stems, merge or miss trees under dense canopy, misjudge stem diameter and therefore root protection areas, and know nothing about tree preservation orders or conservation area status. Under LI Rule 6 and R6g, a landscape architect should involve an arboriculturist where the task goes beyond their competence. If misplaced constraints lead to damaged or felled protected trees, claims and enforcement can follow, and PI usually responds to the professional error, subject to the wording.

How AI is used for tree mapping

Drone photogrammetry, aerial LiDAR and street-level imagery can be processed by AI models that detect individual trees, outline canopies, estimate heights and sometimes predict species. Landscape practices use these outputs to produce early constraints plans, plot root protection areas and feed generative layouts before a full arboricultural survey is commissioned, or sometimes instead of one. The outputs arrive as tidy GIS layers, which makes them look more certain than they are.

The value is real: AI processing can cover a large or inaccessible site quickly and flag where a full survey should concentrate. The problem arises when an indicative layer is passed to a design team, a contractor or a planning authority as if it were a surveyed and categorised tree record, without the limitations travelling with it.

The duties that apply

The LI Code of Practice points to three questions before AI tree data is relied on:

Rule 6 and R6g: is the practice competent to interpret the data, or does it need an arboriculturist to survey and categorise the trees?
R4g(c): does the written agreement make clear whether a tree survey is in scope, who provides it and what limitations apply?
R11g(a): has the client been told the constraints plan is AI-derived and what it can and cannot show?

If the plan later supports appeal evidence, the Planning Inspectorate's AI casework guidance requires intentional AI use to analyse data to be declared, with the checks made.

What can go wrong

AI tree detection has known weak points:

stems misplaced where canopy overhangs, shifting the root protection area;
multi-stem trees, hedgerow trees and understorey missed or merged;
stem diameters estimated from height or canopy, producing the wrong protection radius;
no information on condition, veteran status or protection under a tree preservation order or conservation area;
outdated imagery that predates felling or growth.

Illustrative scenario. A practice issues a constraints plan built from AI analysis of drone data, and the developer's contractor sets out a haul road and compound from it. A protected oak was recorded as two smaller trees with small protection areas. Construction traffic compacts its rooting zone and the tree declines. The council takes enforcement action against the developer, who claims the costs, replacement planting and delay from the practice.

How PI usually responds

A claim that a negligent constraints plan led to tree damage, enforcement and delay is usually a professional negligence claim, and landscape PI commonly responds, subject to the terms. Insurers may ask whether the practice was competent to produce the plan, whether the appointment included a tree survey and whether the client was warned it was not a full survey. Fines and penalties imposed on the developer are generally not insurable, though related costs claimed from the designer may be. Arboricultural consultants carry their own cover; see PI for ecologists and arboriculturists. Physical damage caused directly by site activity is more often a public liability matter.

Practical controls

AI tree data is most useful as a first pass that a specialist then verifies:

label AI-derived tree layers as indicative and not a tree survey;
commission or recommend an arboricultural survey before layouts are fixed;
check protected status with the local planning authority records;
ground-truth stem positions and diameters for trees near proposed works;
carry verified constraints into generative layouts and planting plans;
record the data date and source on every plan.

Similar issues for habitat mapping are discussed in AI habitat mapping from satellite and drone data.

How Apex helps

Apex arranges PI for landscape architects and for arboricultural consultants. A broker can check how your cover treats survey-based advice and where public liability needs to sit alongside it. See the AI in landscape architecture overview and PI and AI errors.

Sources

Frequently asked

Can AI replace an arboricultural survey?

No. AI can detect trees and canopies from drone, LiDAR or aerial data, but it cannot reliably assess species, condition, stem diameter or protected status. LI Rule 6 and R6g point towards involving an arboriculturist where the work is beyond the landscape architect's competence.

Who is liable if a protected tree is damaged after an AI survey?

It depends on the facts. The contractor, developer and designer may all face claims. If the constraints plan was negligent, the developer may claim against the practice, and PI usually responds to that professional error, subject to the policy wording.

Do I need public liability as well as PI?

Many landscape practices hold both. PI usually responds to claims arising from negligent professional advice and design, subject to the policy terms. Public liability typically covers injury or property damage caused by the business's activities, such as staff on site. Where a claim falls depends on how the damage arose and the wordings.

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.