Aipedia: Building control approvers
AI image tools can flag possible missing fire stopping, cavity barriers or insulation gaps in site photos, but silence from the tool is not evidence that the work complies. Inspectors are judged on what a competent inspection should have found.
Part of: AI in building control approvers
In short
Image-recognition tools that scan inspection photos for defects can help inspectors spot issues across large photo sets, but they produce false negatives and false positives. Under the RBI code the inspector must be trained in the technology and manage its risks (2.11), is accountable for decisions (4.2) and must keep auditable records (Annex 1.2). Whether a missed defect is the inspector's negligence or the builder's responsibility depends on the facts, but building control bodies may be joined in claims. PI usually responds to negligence allegations, subject to fire safety restrictions.
Inspection photos now run into the hundreds on a single project, especially where remote inspection, contractor uploads and 360-degree capture are used. Image-recognition tools can sort photos by element, highlight apparent gaps in insulation boards, missing or displaced cavity barriers, unsealed penetrations through compartment walls, or missing fire collars. Some produce a defect list ranked by confidence.
The tools are trained on examples and look for patterns. They have no knowledge of the design, the specified product, or what should be behind a finish. They can confirm that something resembling a cavity barrier is present, but not that it is the right product, correctly fixed and continuous. The same false-assurance problem arises earlier in the process, when rule-checkers review drawings; see AI plans assessment.
Several existing duties apply to this work:
See the RBI code of conduct and operational standards rules. RICS members also need a written reliability decision for outputs they rely on under the RICS AI standard.
False negatives are the main problem. A tool may miss a gap in a cavity barrier hidden by shadow, a penetration behind a pipe, or fire stopping that looks right but uses the wrong material. False positives create a quieter risk: an inspector who learns to dismiss frequent flags may dismiss a real one.
Illustrative: an inspector reviews several hundred photos of a timber-frame residential block using an image tool. The tool flags a handful of insulation gaps, which are put right. It does not flag a run of cavity barriers missing at a floor junction on one elevation, partly obscured in the photos by a scaffold board. Years later a fire spreads within the cavity. The owner and residents bring claims against the builder, the designer and the building control body, alleging the omission was visible in the inspection photos.
Whether that omission is the inspector's negligence or the builder's failure is a question of fact and law. The builder is primarily responsible for compliance, but a building control body that held clear photographic evidence of a defect may struggle to argue it could not reasonably have seen it.
A claim that an inspector missed a visible defect is commonly treated as professional negligence, subject to the policy terms. Where the defect concerns fire spread, check how the policy treats fire safety: construction PI wordings may have restrictions, which are explained on our fire safety PI exposure page. Claims of this kind can arrive long after the work, so run-off and records matter; see AI-written inspection records.
These steps reduce the risk:
See also AI-assisted remote inspections and the AI in building control hub. Apex arranges PI for building control bodies and self-employed inspectors, and a broker can check how your policy treats AI-assisted inspections and fire safety claims.
They can flag possible gaps, but they cannot confirm the correct product, fixing or continuity, and they miss defects hidden by shadow, obstruction or angle. RBI code 2.11 requires inspectors to understand these limits before relying on the tool.
The builder is primarily responsible for compliance, but a building control body may also face a claim if a competent inspector should have spotted the defect. Using AI does not move the responsibility to the software supplier; the inspector remains accountable under RBI code 4.2.
It depends on the wording. A missed defect is usually treated as negligence, but construction PI policies may restrict cover for fire safety or cladding claims. Check the fire safety terms and any AI clauses with your broker.
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 callbackApex 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.