Artificial intelligence has moved into everyday working life faster than almost any technology before it. Most of us now use it without thinking, to draft emails, summarise documents, search for information, write code, and increasingly to support decisions that once demanded real effort. Whether we have chosen it deliberately or not, AI is becoming part of the infrastructure of knowledge work, and fire engineering will be no exception.

The question is no longer whether AI will influence the profession, because it already does. The more useful question is how we integrate it into a discipline whose entire purpose is to protect people and to demonstrate safety in the built environment. That distinction matters, because fire engineering is not simply a technical activity. It is a professional one, carried out under accountability, and the trust placed in us by society does not transfer to a piece of software.

As a profession we have spent years thinking hard about competence, accountability, and ethics, particularly as the industry has come to accept that compliance alone is not enough. Following guidance, applying an accepted solution, or reproducing a previous design may form part of the work, but none of it removes the obligation to understand performance. Behind every fire strategy sits a single question: will this building perform safely in the event of fire? Everything else exists to support answering it.

This is why first-principles knowledge still matters. Fire engineers need to understand fire dynamics, human behaviour, structural response, and smoke movement, along with the assumptions buried inside the tools they use. More than that, they need to understand how those elements interact to deliver a safety objective. A competent engineer does not simply apply a method. They understand why it works, where it applies, and where it does not, and that understanding becomes more important, not less, once artificial intelligence enters the picture.

The opportunity is exploration, not automation

Much of the current conversation about AI fixates on automation, and the anxieties that come with it: which tasks will disappear, which skills will atrophy, whether reliance on a machine will hollow out professional competence. Those concerns are legitimate and deserve serious attention. But they can crowd out a more interesting possibility.

The greatest strength of AI may not be its ability to do work, but its ability to help us explore. Engineering is fundamentally an investigative process. We compare options, test assumptions, examine sensitivities, and try to understand how a system behaves under conditions we cannot fully control. The constraint is rarely a shortage of ideas; far more often it is a shortage of time to pursue them. A fire engineer under programme pressure might run one smoke-control scenario when the design really warranted five, or accept a single design fire rather than testing how the result moves as the growth rate changes.

This is where AI earns its place. It can generate alternative approaches, reframe a problem, draft a calculation, write a script, and summarise guidance far faster than we can by hand, which lets the engineer spend less time on mechanical work and more time thinking critically about performance. The article you are reading is a small example of the same principle. We used AI while drafting it, not because the software understands fire engineering, and not because it produced the ideas, but because the audience, the objective, and the key messages were defined first, and AI was then used to explore wording and structure around them. When the intent is clear and the objective is well defined, AI can widen the space of options you consider and surface approaches you might otherwise have missed. The engineer still defines the problem, evaluates the alternatives, and decides which one is appropriate.

The same pattern holds in design. AI can help investigate a strategy, compare options, interrogate assumptions, and even build the tools used to do those things. What it cannot do is take responsibility for the outcome.

Plausible is not the same as correct

One of the real risks of AI is that its answers look right. They arrive fluent, confident, and well formatted, which makes them easy to accept and hard to challenge. This is not a new problem. Fire engineers have always had to be careful with spreadsheets, software packages, and models, because the danger was never the existence of a tool; it was the uncritical acceptance of its output. A view-factor calculation to BR 187 returns a number whether or not the geometry was set up correctly, and a CFD model renders a convincing plume whether or not the boundary conditions make sense.

AI simply sharpens a lesson the profession already knew: trust has to be earned through understanding. That makes transparency more valuable than ever. Methods should be visible, assumptions should be explicit, limitations should be understood, calculations should be traceable, and any output should be open to challenge by another competent person. None of these principles are new. If anything, AI reminds us why they were worth insisting on in the first place.

Where OpenFire fits

This is the gap OpenFire was built to close. The principles above are easy to state and hard to honour when the tool itself is opaque, because an engineer cannot interrogate what the software will not show them. A calculator that returns only a number quietly encourages the profession to trust the number, which is exactly the habit AI makes dangerous.

OpenFire takes the opposite approach. Because its methods are open, the calculation behind a result is there to be read rather than sealed inside the software. An engineer can follow the method itself, check the assumptions it relies on, see the range of inputs over which it remains valid, and watch the answer move as those inputs change. Transparency, traceability, and the ability to challenge a result stop being aspirations and become the way the tool works. That is the foundation on which AI assistance can be added responsibly, because a recommendation is only as trustworthy as the method a competent engineer can check it against.

You can read through the open method library, including the external fire-spread view-factor method from BR 187, at the OpenFire method catalogue; it is open to browse without an account.

A profession still anchored in judgement

The best fire engineers have always been creative. They question assumptions, spot opportunities, and find solutions that satisfy a safety objective while serving the architectural, operational, environmental, and commercial goals around it. Used well, AI can support that creativity: less time spent searching for information and more spent understanding it, less time reproducing old solutions and more spent exploring new ones.

These questions run deeper than any single article can. We look more closely at what responsible AI means for regulation and approval, for fire modelling and the black box problem, for how the next generation of engineers learns the profession, and for what it demands of those who lead engineering teams.

So the real choice facing the profession is not whether to accept or reject AI. It is whether we integrate it consciously, transparently, and responsibly. As with every significant technology before it, the outcome will depend far less on the tool than on the competence, judgement, and integrity of the people using it. Technology will keep advancing at remarkable speed. The responsibility to protect people will not move, and it remains, as it always has, in the hands of the engineer.