There’s something curious about our relationship with technology. The better it works, the less we think about it.
We board trains, travel by plane, undergo medical procedures involving robotics, and even trust banking apps with our savings, without really stopping to wonder what’s going on behind all those screens, sensors, and systems that, in theory, are there to make our lives easier.
We create technology to make life easier, reduce our mistakes, help us go further, and, perhaps most importantly, protect us from ourselves.
But there’s quite a big difference between trusting something and assuming it will never fail. And I have a feeling that, somewhere along the way, we started confusing the two.
Two accidents and far too many questions.
2026 has left us with images that are difficult to forget, particularly for those of us in Spain. First, Adamuz, in January. And, more recently, Urda, just a few days ago.
Two tragedies that have made us question railway safety, infrastructure, procedures, and the responsibilities of those involved in keeping these systems running.
We still don’t know exactly what went wrong, which circumstances came together, or what decisions may have influenced the outcome of each accident. And obviously, I’m not about to start speculating. That’s what the investigation teams are there for, and I’m pretty sure the professionals involved don’t need another person throwing their own theories into the mix.
But there is something these events have made me question, something that I believe goes far beyond the railways themselves.
What happens when we place so much of our safety in technological systems we’ve come to take for granted, and suddenly, something stops working when we least expect it? Are we still capable of reacting? Can we recognize when technology is failing? And, perhaps most importantly, who decides when we can no longer trust it?
Simple: technology works. Until it doesn’t.
When was the last time you drove somewhere unfamiliar without using Google Maps? Personally, I can’t even remember.
We’ve become so used to following a blue line on a screen that we no longer make much effort to remember the route. Half the time, we’re not even fully aware of where we’re going. If GPS disappeared tomorrow, that old joke about how you’re not a true madrileño until you’ve accidentally driven all the way around Madrid’s M-30 ring road would probably make a comeback.
Now, this isn’t exactly a tragedy, but it’s an interesting example of how our relationship with a particular skill changes when we stop using it.
Automation reduces workloads, improves accuracy, and can help prevent human error. But it can also create a dependency that, if not properly managed, weakens our ability to react.
There’s actually a name for this: automation complacency, a rather fancy term I only recently came across. It’s our tendency to trust an automated system so much that we stop paying the same level of attention to what it’s doing.
And here’s where things get a little uncomfortable. The better a technology works, the less we need to intervene. But the less we intervene, the fewer opportunities we have to prepare ourselves for when it fails.
So here’s what I’m wondering. If one day something goes wrong, alarms start going off, and suddenly someone has to make a critical decision in a matter of seconds, are they actually prepared?
Because having a human in control is one thing. Having a human who’s ready to step in when things go wrong is something else entirely.
Technology ages too, even if we don’t like talking about it.
We’re obsessed with innovation, and we love imagining what the future might look like. The problem is that we’re not quite so keen on admitting that a significant part of our essential infrastructure still relies on technologies developed decades ago.
And before anyone throws their hands up in horror, let’s get one thing straight.
Just because a technology is old doesn’t mean it’s unsafe.
Some systems have been running for decades precisely because they’ve proven to be incredibly reliable. The problem isn’t getting older. It’s when we stop questioning whether a technology is still suitable for the environment in which it operates.
Components age, manufacturers disappear, professionals who understand certain systems retire, and needs change. And I’d be naive not to acknowledge that upgrading certain infrastructure can involve investments worth millions, operational disruptions, and new risks associated with the transition.
Take banking. Much of the world’s core banking infrastructure still runs on COBOL, a programming language created in 1959. It quietly processes enormous volumes of transactions every day, from ATM withdrawals to card payments, and it does so reliably.
The risk isn’t the language itself. It’s that the people who truly understand these systems are retiring, the documentation is often thin, and decades of patches have made them so intertwined that replacing them feels almost as dangerous as keeping them.
But what happens when the cost of modernization becomes the main argument for continuing to use something that might actually need replacing? And who’s assessing the risk of doing nothing?
I genuinely think this is a conversation we should be having much more often. Because we can put the cost of upgrading infrastructure into a budget, but the cost of not doing so, as Adamuz and Urda have reminded us, can sometimes be far too high.
Maybe we don’t need to trust technology less. We just need to trust it differently.
For years, we’ve talked about human error as something technology should help us minimize as much as possible. And to a certain extent, that makes sense. We get tired, distracted, forget things, misinterpret information, and make rushed decisions.
We are human. Get over it.
But there’s something that makes me wonder whether we’re approaching our relationship with technology in quite the right way. Are we using automation to complement our abilities, or are we starting to blindly replace them?
Because when something goes wrong, we often expect a person to step in. To understand the situation, identify the problem, and make the right decision. And here’s the contradiction I can’t quite get past: we can’t keep delegating more and more to technology while expecting humans to retain exactly the same ability to react when it fails.
Having someone sitting in front of a screen isn’t enough to call it human oversight. We need training, simulations, clear procedures, and, above all, people who genuinely understand the systems they’re working with and their limitations.
And that brings me to an even bigger question. We’re entering an era where we’ll be handing over increasingly important decisions to artificial intelligence. And no, I don’t think we should suddenly start distrusting everything or go back to manual systems out of nostalgia.
I simply think we need to learn to trust differently. To understand what technology can do, what it can’t, and what we need to know ourselves to remain in control. Because safety shouldn’t depend on technology never failing, but on our ability to anticipate, manage, and recover when it does.
And perhaps the question we should be asking more often isn’t how far technology can go, but whether we’re actually prepared to keep up with it.
I don’t have a simple, straightforward answer to everything I’ve raised here. But these recent events have reminded me of something important: technology can help us build a safer world, but it should never become an excuse to stop understanding it.
I just hope that, if we ever need to take back control, we still know how.