FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Structural and civil engineers
IStructE advises keeping the most junior engineers away from AI tools until they can show the underlying competence, so supervisors must decide who uses AI and check what they produce. A graduate who cannot spot a wrong answer cannot safely use a tool that produces convincing ones.
Part of: AI in structural and civil engineering
In short
Graduate engineers learn by doing calculations, checking load paths and having their work reviewed. AI can short-circuit that learning and produce work that looks better than the graduate's actual understanding. IStructE's AI guidance says firms should keep the most junior engineers away from AI tools until they demonstrate competence and should train staff. The Statement of Ethical Principles allows work under competent supervision, and Part 2A recognises supervised trainees. If a supervised graduate's AI-assisted work goes wrong, the firm is usually judged against the standard of a reasonably competent engineer, and PI usually responds subject to the wording.
Experienced engineers use instinct built over years to sense when a number is wrong. Graduates are still building that instinct. A generative AI chatbot or calculation tool that produces a tidy answer can hide an error the graduate does not have the experience to spot, and repeated reliance can slow the development of the judgement the graduate needs. CROSS-UK Report 1554 described an AI-written calculation report that looked robust to a non-engineer but was fundamentally wrong. A graduate using AI without close supervision is in a similar position.
The IStructE and Engineering Council guidance explainer sets these out in full.
In negligence claims, engineers are usually judged against the Bolam standard, derived from Bolam v Friern Hospital: whether the work met the standard of a reasonably competent member of the profession. A client is unlikely to accept a lower standard because a graduate did the work. The question will usually be whether the firm's supervision and checking brought the output up to that standard. Where AI was used without review, that is hard to show.
Illustrative scenario: a graduate uses an AI tool to design a set of lintels for a refurbishment. The supervisor, under time pressure, signs off the drawing without reviewing the calculations. One lintel is undersized for a point load. Cracking follows, and the client claims remedial costs. The firm's records show no training on the tool and no review of the graduate's work.
PI usually responds to claims arising from work by employees, including graduates, as part of the firm's professional business, subject to the policy wording. Insurers may ask about supervision and quality management at renewal. Records of training, competence sign-off and review strengthen both the defence of a claim and the firm's presentation to insurers.
Supervision failures can also widen a claim. If one graduate used an AI tool without review on several jobs, the same error may repeat across projects, and the firm may face a series of related claims. How the policy aggregates related claims, and whether the limit is enough for that scenario, are worth discussing with a broker.
Related: LLMs and structural calculations, AI design checking and PI for new engineers. Back to AI in structural and civil engineering.
Apex arranges PI for engineering practices of all sizes. A broker reviews your cover and how your proposal form describes supervision and AI use by junior staff.
IStructE advises keeping the most junior engineers away from AI tools until they can show the underlying competence. That is a gate, not a permanent ban. Once a graduate can show competence in a task, supervised AI use for that task may be appropriate.
Usually not. The standard is generally that of a reasonably competent engineer doing that work. The firm's supervision and checking need to bring a graduate's output up to that standard before it is issued.
The firm is responsible to the client, and the supervising engineer is professionally accountable for the review. Article 5 of the IStructE Code of Conduct asks engineers who oversee others to communicate the limits of their own knowledge, so supervisors should also know the tools well enough to review AI output.
Records of training on each approved tool, competence sign-off for tasks before AI use, supervisory reviews of AI-assisted work and any concerns raised. These support Code of Conduct Article 4 and help demonstrate reasonable supervision if a claim arises.
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.
