Summary
- A competent HSE professional using AI well can outperform the same professional without it, but only if they stay in charge of the reasoning.
- The biggest gain is not speed. Used as a sparring partner that challenges your work, AI can make you better at the job.
- The same tool can build competence or erode it. “Challenge my risk assessment” and “write my risk assessment” lead to very different places.
- The documents look the same either way, so leaders need to watch how their people work, not just what they produce.
I have worked in various aspects of HSE since the late 1980s, and I have lived through a great many changes in that time. None, however, comes close to what we are witnessing now. AI is more than a game changer. It is a paradigm shift, and it will affect almost everyone, particularly those whose work rests on expertise and judgement rather than routine, i.e., the HSE profession.
In this series, AI and HSE Management: The Good, the Bad and the Accountable, I want to share my thoughts on what I see as one of the central issues for the HSE profession: AI can make HSE professionals markedly better or quietly worse, depending on how it is used. Either way, it does not change who is accountable, and GCC regulators are already writing that principle down.
In my previous AI series AI and the HSSE Profession: What’s Actually Changing, I argued that accountability stays with the organisation and the professional, that an AI tool is only as reliable as the sources it is grounded in, and that the answer is an expert-in-the-loop model operating within a governed, auditable workflow. I stand by all of it.
But that argument rested on an assumption: that the expert in the loop remains an expert. This series tests that assumption. It asks what AI does over time to the people who use it, to the independence of the checks we rely on, and to the information we entrust to it, and what regulators and boards now expect in return.
We will get to the harder questions in later posts. But let me start on a positive note, because the positive case is real and it is substantial.
The argument in one sentence
A competent HSE professional using AI well can outperform the same professional without it.
Note the conditions: “competent“, and “using it well“. The gain does not come from the machine doing the thinking. It comes from the machine removing much of the low-value cognitive work that has always surrounded the thinking, and, used properly, from the machine sharpening the thinking itself.
More than faster
The productivity case is familiar by now. AI can find the requirements that apply to an activity across six GCC jurisdictions and their free zones, compare a national regulation against an ISO standard and a client specification, turn a pile of witness statements and permit records into a coherent timeline, check a procedure for omissions, spot where a method statement and a risk assessment disagree, surface past incidents that resemble tomorrow’s job, and map compliance evidence to obligations before the auditor arrives.
None of that is trivial, but none of it is where an HSE professional adds real value. That lies in judgement, and the tasks above are the scaffolding around it. Let AI carry more of that scaffolding and a good professional has more time on site, more time with supervisors, and more time asking whether the controls on paper actually work in practice
But there is a second benefit that gets far less attention. Used the right way, AI does not just make professionals quicker. It can make them better.
AI as a sparring partner
Every good HSE professional I have worked with had someone early in their career who challenged their work: a senior engineer who asked “what happens if that isolation fails?“, or an auditor who would not accept “operator error” as a root cause. That kind of challenge is how judgement is built. It has always been in short supply.
AI can supply some of it, on demand, if we ask it to.
Challenge the assessment. Rather than asking AI to write a risk assessment, write your own and ask it to attack it. Assume the key control fails: what is the most likely way it fails, and what then?
Look beyond your own experience. Every professional has blind spots shaped by the industries and sites they know. Asking “what hazards might I have missed for this activity?” draws on a far wider base of experience than any one career.
Test the investigation. Compare your findings against recognised causal models, such as bow-tie analysis or human factors frameworks, and ask whether you stopped too early. Many investigations end at the person closest to the event. AI is a useful prompt to keep asking why.
Learn from other people’s incidents. The lesson that would have prevented your next incident has often already been learned somewhere else. AI can help find it.
Ask for the reasoning, not just the answer. A junior adviser who asks why a requirement applies, and what would change if the facts were different, learns something. One who asks only what the requirement is learns very little.
In each case the professional does the reasoning and the AI tests it. That is the difference that matters.
The same tool can be used in two very different ways
| AI augmentation | AI substitution |
| Find the applicable requirements, cite the legal reference for each, and flag where applicability is uncertain. | Tell me the requirements and I’ll accept them. |
| Challenge my risk assessment. | Write my risk assessment.” |
| Challenge my risk assessment. | Identify the hazards for me |
| Compare my investigation against these causal models. | Determine the root cause. |
| Find weaknesses in my proposed controls. | Tell me what controls to implement. |
In the left-hand column the professional stays in charge of the reasoning. The AI searches, challenges, compares and probes, and the human decides. Every challenge it raises is something the professional must think through and answer, which is precisely how competence grows.
In the right-hand column the reasoning has been handed over. The output may look just as good, sometimes better. But the professional is no longer exercising the skills that make them worth employing: observing, interpreting, challenging and judging.
The first column builds competence. The second can erode it.
How do you tell which column your team is working in?
This is the question I would put to every HSE leader, and it is harder to answer than it looks, because the documents produced in both columns can be indistinguishable. A few signs are worth watching for.
Where did the reasoning start? Compiling structured information from a validated source, such as a first draft of a legal register, is legitimate AI work, provided a competent professional reviews it. Judgement documents are different. If AI frames the risk assessment, the root cause or the choice of controls before the professional has thought about it, the reasoning has already been set by the machine, and review tends to become editing.
Can the author defend it without the tool? Ask the person who signed a risk assessment why a particular control was chosen over the alternatives. Someone working in the left-hand column will tell you. Someone working in the right-hand column may simply tell you what the document says.
Does the document look like the site? Generic hazards, generic controls and generic wording across different sites and different jobs suggest the output is not being tested against real conditions.
What was changed at review? If AI output routinely passes review unchanged, either the tool is remarkably good or nobody is challenging it. In HSE, the second is more likely.
Are your junior people still learning? Experienced professionals can generally use AI safely because they know what good looks like. People early in their careers may never develop that sense if the machine has always supplied the answer.
None of these is a test on its own. Together they tell you a good deal, and none of them is an IT question. They are questions for line managers and HSE leaders about how their people are working.
A positive start, with a condition attached
I am optimistic about what AI can do for the HSE profession. Used well, it will make good professionals considerably better and give them back time for the work that keeps people safe. But “used well” carries a lot of weight in that sentence. The benefit belongs to the professional who stays in charge of the reasoning, and it depends on organisations deliberately keeping their people in the left-hand column.
In my previous series I described the false authority of AI output: how it can look complete and correct when it is neither. In the next post I look at what living with that does to people over time, and how AI can raise the apparent quality of HSE documentation while quietly lowering the competence behind it.
Why Redlog Is Building for This Future
At Redlog, we are building compliance intelligence tools on the principles set out in my previous series: curated and expert-validated regulatory content, jurisdiction-specific coverage, source traceability, applicability-aware outputs, and an expert-in-the-loop model that keeps professional accountability where it belongs.
This series adds a further principle. AI should strengthen the professional’s judgement, not stand in for it. Tools that show their sources and flag their uncertainty prompt the professional to check, challenge and decide. That keeps them in the left-hand column.
We are publishing this series because the conversation about AI and accountability in HSE should happen inside the profession. It should be shaped by people who understand regulatory compliance, duty of care, operational risk and the consequences of getting this work wrong.
AI will change HSE practice. That is no longer the question. The question is whether the profession will shape that change with judgement, governance and accountability, or allow it to be shaped elsewhere.
If that is how you think AI should work in HSE compliance, it is the conversation we want to have, starting with our webinar.
This is Part 1 of AI and HSE Management: The Good, the Bad and the Accountable. The previous series, AI and the HSSE Profession: What’s Actually Changing, and all of our prior blog posts are available at www.redlogenv.com/blog.
- The Competent HSE Professional, Amplified - September 29, 2026
- Confused About Heat Stress and the GCC Midday Work Ban? - June 27, 2026
- The HSSE Professionals Who Shape AI Will Define the Next Decade - June 19, 2026