The statement, a 42-page document published this month, was written by an AI working group drawing on building engineering, digital engineering, academia and regulatory compliance, then reviewed by CIBSE’s Technical Policy Committee and discussed with its Society of Digital Engineering. Ahmed Hegazi, CIBSE’s head of knowledge and digital, said the institution wanted to get ahead of a shift already under way in how buildings are designed, analysed and run. “CIBSE has always been at the forefront of building engineering, and our approach to AI is no different,” he said.

The statement takes what CIBSE calls an opportunity-first position. It does not discourage engineers from using AI tools and describes AI as a way to extend what engineers can do rather than replace them. Hegazi put the limit on that opportunity plainly: “technology must support professional judgement, never replace it.”

That line carries weight because of what sits underneath it. Whether or not an engineer used an AI tool to reach a conclusion, CIBSE’s position is that the engineer remains professionally and legally responsible for the result. Outputs from AI systems, including large language models and the newer category of agentic AI that can take actions rather than just generate text, have to be checked and verified by a qualified person before anyone relies on them. The statement sets out that the level of oversight should scale with how much is riding on the decision: a tool helping with a routine calculation needs less scrutiny than one feeding into a judgement about a safety-critical system.

The timing is not incidental. CIBSE is explicit about where the statement sits alongside the Building Safety Act, at a point when the Building Safety Regulator is already pressing for clarity on who is accountable for decisions on higher-risk buildings. The position closes off one excuse before it gets used: an AI recommendation does not dilute, transfer or excuse an engineer’s own duty to understand and stand behind the advice they give.

The document also covers ground that building services engineers haven’t traditionally had to think much about. It addresses confidentiality and data protection when feeding building data into third-party AI tools, intellectual property, contractual liability, insurance, the environmental cost of running AI systems, and what ongoing training should look like for engineers expected to use these tools competently rather than simply operate them.

CIBSE has been clear that this is not the final word. It describes the statement as a living document that will be reviewed as AI tools and regulation develop, and is inviting feedback from members and the wider industry. The AI working group that produced it is expected to keep shaping the position as practice moves on.

For facilities teams, the practical consequence sits less with the engineers who produced the statement than with the people buying the tools engineers increasingly rely on. Building management systems, predictive maintenance platforms and compliance auditing tools are being sold with AI features built in, often with limited detail on how outputs are checked or who is accountable if something goes wrong. CIBSE’s statement gives FM and procurement teams a reference point they didn’t have a fortnight ago: a professional body’s own description of what responsible AI use should look like, and grounds to ask vendors harder questions about verification and accountability before signing a contract.

Register your interest for The Workplace Event | The Facilities Event 2027 

Subscribe to Workplace Unplugged for monthly updates!