Trust is a must

There’s a moment I hear about constantly from EHS professionals. A frontline worker spots a hazard. They pull out their phone, open the reporting app and then they stop. They put their phone away. They don’t report.

It’s not apathy. It’s something more specific: the last time they raised a concern, nothing happened. No acknowledgement, no follow-up, no visible action. So why would this time be any different?

That’s not a technology failure. It’s a failure in trust. And no amount of software investment fixes it on its own.

As AI accelerates EHS capability, the temptation is to focus on features from smarter hazard recognition and predictive analytics to automated workflows. These things genuinely matter. But organisations that lead with features and skip the trust conversation are building on sand.

The psychological barrier

When workers encounter a new EHS system, two things happen almost simultaneously. They assess whether it’s easy to use. And they assess what it’s actually for. Is this to help me do my job safely or to monitor me?

The surveillance perception is real, and it’s rational. Workers in high-risk industries have watched technology arrive in their workplaces under the banner of safety, only to find it used for performance management or blame attribution. Once that association forms, no amount of training or change management messaging will shift it easily. You have lost hearts and minds.

The design of the system matters as much as the intent behind it. If the first thing a worker encounters is a form requiring extensive detail with no obvious benefit to them personally, no feedback, no acknowledgement, no visible outcome, they read the signal clearly. This is more admin, not support.

I started my career in cybersecurity before moving into workplace safety, and the parallels are striking. Security teams once faced the same challenge with breach reporting, employees who saw something suspicious, wouldn’t say anything because they feared blame, or because previous reports went nowhere. The organisations that cracked it didn’t do so by building better reporting systems. They did it by demonstrating, consistently and visibly, that reporting led to action. The technology then became the vehicle for positive change

The data cost of ease

There’s a direct link between system usability and the quality of safety data an organisation can act on. Sadly a lot of organisations haven’t drawn it.

If a reporting process is clunky, time-consuming, or doesn’t work well on a mobile device, people won’t use it, full stop. That means the data your safety team is working with is skewed, missing any low-level signals that are often predictors of something that needs addressed. 

I’ve seen organisations sit on years of incident data that looks reassuringly thin, only to discover when they implement a better system that reporting volume doubles almost immediately. The incidents were always happening. They just weren’t being captured. What looked like a good safety record was actually a data gap.

This matters enormously as organisations look to AI and analytics to surface patterns and predict risk. The insight is only as good as the input. Garbage in, garbage out – except in safety, the consequences of bad data aren’t just analytical errors. They’re real people getting hurt.

Ease-of-use isn’t a nicety. It’s a safety-critical requirement.

Closing the loop

The most practical change management intervention I’ve seen organisations make is also one of the simplest: tell people what happened next.

When a worker raises an issue and then receives, through the system, or in person, or both, some form of acknowledgement that it was received, reviewed and acted upon, something shifts. The next report comes faster. Then another one. Reporting culture builds on itself, but only if the feedback loop is closed.

Many organisations once did this with nothing more than a noticeboard – a simple “You Said, We Did” board where raised concerns sat alongside the actions taken in response.  Low-tech, but effective, because it made the loop visible to everyone. The principle hasn’t changed.  What’s changed is our ability to do it at scale.

This is where technology can play an enabling role that’s genuinely meaningful, not automating human judgement, but making it visible. Automated notifications when an action is assigned. Status updates when a corrective measure is completed. Simple dashboards that show frontline teams what’s been raised, what’s in progress, what’s resolved. None of this requires sophisticated AI. It requires intentional design

AI as accelerator, not replacement

AI is already in the picture and embedding itself fast. The same trust principles apply.

The most dangerous framing is AI as a replacement for human safety judgement. Workers and safety professionals alike are right to be sceptical of systems that claim to make decisions on their behalf. The moment it feels like the technology is running the programme rather than supporting the people who run it, trust collapses. We are left with a 1984 Big Brother culture.

What I’ve seen work is AI positioned as a strategic lever, something that does the analytical heavy lifting that used to consume hours of a safety professional’s week, and surfaces it as a recommendation rather than a directive. The night shift has 30% more incidents involving a particular piece of equipment. Here’s what the data suggests. Now, what do you want to do about it?

That framing keeps the human in the feedback loop, which is where they belong. It makes AI feel like a trusted advisor rather than an overseer. And critically, it nurtures trust in the technology, because people see it working for them rather than to them.

The organisations getting EHS technology right aren’t necessarily the ones with the most advanced systems. They’re the ones that treat trust as foundational, not as an afterthought.

Jose Arcilla is CEO at HSI. For more information, visit www.hsi.com