Construction professionals should expect to challenge and verify AI outputs – and the professional institutions should develop the guardrails around the use of AI.
That’s one of the key messages in a new report from the Built Environment Futures Assembly (BEFA) – the University of the Built Environment leadership forum chaired by Modernise Or Die author Mark Farmer – and Harlow Consulting.
The report states that AI-assisted outputs should be treated as drafts or decision support, not authoritative conclusions. “Professionals should be expected to use, challenge, verify, document and then decide, particularly where outputs may influence design, planning, cost, procurement, compliance, safety, carbon or asset performance.
“Responsible adoption will also depend on proportionate assurance. Organisations need practical guardrails covering approved tools, sensitive data, human review, audit trails, disclosure, sign-off responsibilities and escalation where uncertainty remains. These norms should be developed with professional bodies, employers, insurers, clients and regulators, so that responsible innovation is supported without creating unnecessary bureaucracy.”
The report emphasises the role that procurement teams and the insurance industry should take: “Procurement will be an important lever for responsible use, with clients and procurement teams increasingly needing to understand not only whether suppliers are using AI, but how they are using it, with what data, under what controls and with what assurance. Procurement should reward transparency, accountability, evidence of checking and good information management, rather than vague claims about AI capability.”
Mandatory use of AI?
Widespread adoption of best practice in the use of AI may have further impacts, according to the BEFA report. “There is potential transformation of procurement and assurance if AI becomes demonstrably safer and more reliable. In which case, insurers and clients might mandate the use of AI, moving away from the current caution being shown by the insurance sector to the use of such tools during their nascent development and proof of better assurance,” it says.
“In particular, the parameters of AI use that professional indemnity insurers are prepared to underwrite will also very likely become the ‘guardrails’ that guide AI adoption.”
The report argues that AI will affect value, commercial models and long-term sector outcomes. “It may reduce the time required for some professional tasks, especially document production, analysis and routine processing. This will challenge business models based heavily on time spent rather than value delivered, and raise questions about pricing, liability, training, client expectations and the role of junior staff.”
