The world of health & safety may not be renowned for innovation, says Cam Stevens, but that’s about to change due to a convergence of AI, automation and safety technologies.
Safety is not renowned for innovation, and it’s not hard to understand why.
Our discipline is built around reducing uncertainty. We are trained to anticipate what could go wrong and then do something to help make sure that doesn’t happen: strengthen controls, develop safe work procedures, design out the exposure, or make sure the right conversations happen before the work starts.
Risk is the effect of uncertainty on objectives, and most of our professional attention goes into shrinking the downside of that uncertainty.
Innovation asks something different. It asks us to tolerate a level of uncertainty for long enough to find out whether something new might create value. That generally doesn’t sit easily in a profession trained to protect organisations from harm and which is suspicious of untested claims, especially when those claims come in the form of a hyped-up product demo, a glossy slide deck and a vendor promising to transform safety.
It’s okay to proceed with caution. A safety professional should be wary of an unproven claim, and should ask what problem is being solved, what evidence exists, who carries the risk, what changes for workers, and what happens when it fails. These are great questions, and we should keep asking them.
The tension is that uncertainty has two sides. A threat to be mitigated but potentially an opportunity to do things better. If we treat every uncertainty as something to eliminate, we miss the possibility that something new might help us improve things, maybe even transform an industry.
Defining safety innovation
A clear definition helps focus on creating value from the upside of uncertainty. I define Safety Innovation as “the disciplined practice of applying new thinking, methods and technologies to measurably improve the design, experience and safety of work”.
Disciplined, because innovation in safety cannot just mean trying things because they’re interesting. Applying, because an idea that never leaves the lab is not yet an innovation. Thinking, methods and technologies are all there because innovation includes things well beyond technology. And the measure of improvement is whether the work itself is better.
Innovation in safety must connect back to better work. It should make work easier to do well. It should support the people doing it. It should make controls stronger and decisions clearer.
That last point deserves attention, because in safety we can be too quick to define value through the absence of incidents alone. Incident reduction matters, but it is a narrow and lagging measure, and often not where value first appears.
So, what is innovation for better work? Innovation might make a complex procedure easier to use in the field. It might help a supervisor notice a weak signal earlier. It might give workers the information they need at the point of decision or take enough administrative load off a team that a safety professional can spend more time understanding how work is done. None of that is trivial, even if the result isn’t a reduction in a lagging indicator.
It is also why innovation reaches far beyond emerging technology. A better way of running a pre-start conversation can be an innovation. So can a new method for highlighting critical controls within dense procedures, a clearer way of helping leaders to better understand the risks in their business unit, or a more useful approach to learning from everyday work. None of those need a VR headset or a machine learning algorithm to be innovative.
Why technology, and why now?
So why does so much of the current conversation come back to technology? Because, whether we like it or not, emerging tech is where a great deal of genuine novelty has arrived in the modern workplace.
For most of my career I’ve been working closely with AI, automation and emerging safety technologies, partly through my own practice and partly because the leading edge is moving too quickly to ignore. Over the past few years, and particularly the past 18 months, I’ve observed the most compelling change in the tech world is the convergence of major improvements at the same time. Several capabilities are maturing at once, and no single one of them transforms health and safety on its own. We now have systems that can read, summarise, classify, generate, see, listen, detect patterns and increasingly act across workflows. That convergence does not automatically make any of it useful, but it does give us something genuinely novel and different to work with.
From digital tools to learning loops
This is where I think the most interesting shift is taking place. For most of the digital era, we’ve used hardware and software systems as tools. They stored information, moved it around, made it searchable and made it easier to report. Our workers carried the judgment, the context and the interpretation, and the system mostly kept the record.
What is emerging now is closer to a learning loop between people and machines. With modern data architecture and agentic artificial intelligence technologies, for the first time our systems can learn from the judgment of the people using them and begin to improve through an iterative cycle of feedback loops.
In safety, these feedback loops will require the practitioner to integrate their judgment, ethics, nuanced operational context and the read of the room that technology simply can’t learn alone. The machine brings pattern, recall, speed and reach across far more information than any one person could take in. Each can make the other better, if the relationship is designed that way. The potential to drastically change health and safety is there for us to pressure test.
That is a very different proposition from replacing safety professionals with AI, which is both an overstatement and a distraction. The more useful question is how safety professionals use these tools to become better at the parts of the role that matter most: understanding work, engaging people, influencing decisions, improving controls and helping organisations learn.
Going backwards?
There is a line that gets thrown around at conferences: if you are not innovating, you are going backwards. I think it is nearly right. Standing still is not always going backwards, and sometimes it is the sensible thing to do. But the ground is moving. Work is changing, technology is changing, organisations are changing, and what people expect of safety is changing with them. If our practice does not keep learning and adapting, the gap between the safety function and the actual work will continue to grow.
That does not mean every team needs an innovation lab, or that every new tool deserves a pilot. It means having enough curiosity to engage with what is changing, enough discipline to tell substance from theatre, and enough confidence to start with the problem rather than the solution.
From resistance to discernment
For the practitioner who is wary of all this, I would not suggest dropping your guard. A bit of tension is healthy. I suggest shifting from resistance to discernment. You don’t need to become a technologist. You do need enough digital literacy and systems thinking to ask great questions. What problem are we solving? What part of the work will this change? Who benefits, and who might be exposed? What evidence would convince us, and what would make us stop?
Innovation in safety must connect back to better work. It should make work easier to do well. It should support the people doing it. It should make controls stronger and decisions clearer. And if it does not improve the design, experience or safety of work in a meaningful way, it probably doesn’t deserve to be called safety innovation.
Safety may not be renowned for innovation yet. But if we define it properly, apply it responsibly and focus on the design of better work, I’m confident that can change.
This article was published in the July / August 2026 edition of Safeguard Magazine.
