When talking about safety, risk is often imagined as something obvious: a serious failure, an accident, a collapse, a glaring error, a situation that is already out of control. It is an understandable view, but an incomplete one. In reality, many serious problems do not begin as clear emergencies. They begin much earlier, in a reduced, ambiguous, almost negligible form.

A strange smell that comes back from time to time. An unusual noise. A procedure bypassed “just for convenience”. An anomaly that does not yet stop the system but alters its behaviour. A small deviation that, by itself, seems irrelevant. It is precisely here that the issue of weak signals comes into play.

In the language of strategic foresight and horizon scanning, weak signals are early, low-visibility signs of potentially important change. They are not definitive proof. They are not clear alarms. They are early signals, often difficult to interpret, which can nevertheless anticipate larger future developments. This is exactly why they are studied in contexts where anticipating matters more than reacting late.

The problem is not seeing them, but recognising them

The difficulty of weak signals lies not so much in their existence as in the way they are perceived. They often pass in front of us without being read as relevant. The reason is simple: they are small, isolated, and not yet accompanied by obvious damage. As long as nothing actually breaks, the natural tendency is to treat them as background noise.

This is where prevention differs from simple reaction. Those who work well on safety do not wait for the risk to become obvious. They try to understand whether a small anomaly is really just a marginal episode or the start of a more problematic trajectory.

This does not mean living in a permanent state of alarm or turning every detail into a danger. It means doing something more difficult: maintaining enough clarity not to ignore small signals simply because they are not yet large enough to demand attention on their own.

Why they matter so much in safety

In complex contexts, accidents rarely arise from a single sudden event completely separate from everything else. More often, they are the result of a sequence: conditions degrade, margins shrink, habits become established, controls loosen, signals accumulate without ever really being read.

This is why the concept of weak signals is particularly useful when talking about safety. It helps shift attention from the final moment of the problem to its initial stage, when the risk is still more manageable. Horizon scanning is used precisely for this purpose: identifying early signs of potentially important future developments, in order to reduce uncertainty and increase preparedness.

In other words, safety does not depend only on the ability to intervene well when something goes wrong. It also depends on the ability to read in time what is slowly changing before the situation becomes critical.

At home, at work, in infrastructure

The strength of this concept lies in the fact that it applies across very different contexts.

At home, a weak signal may be a socket that gets hotter than usual, a stove giving off an unusual smell, a window that no longer closes properly, what seems to be a minor leak, or a boiler that changes its noise. None of these elements, taken in isolation, automatically amounts to a serious danger. But systematically ignoring them means allowing risk to develop unchecked.

At work, weak signals often take an organisational form: small deviations, near misses, repeated errors, devices used incorrectly, procedures only partly followed, postponed maintenance, informal adaptations considered normal. The problem is that, when these deviations stabilise, they stop looking like exceptions and become operational culture.

In infrastructure, the issue is even more evident. Progressive degradation, minor anomalies, out-of-range instrument readings, performance that slowly declines, operating conditions that worsen without yet becoming dramatic: all of these are examples of fragility that often emerge long before a serious event. The point is not only to detect them, but to avoid interpreting them as irrelevant simply because they have not yet produced a crisis.

The real mistake: waiting for the strong signal

One of the most common mistakes in risk management is thinking that attention should only be activated once the problem is already clear. But the strong signal comes late. When a smell becomes smoke, when a deviation becomes a failure, when a loss of control becomes an accident, the room for correction is already much narrower.

Weak signals, by contrast, have value precisely because they come earlier. They are less comfortable to read, less spectacular, less certain. But that is exactly where prevention has more room to work. The point, therefore, is not to pretend to predict everything, but to develop a more intelligent attention toward what changes slowly, what repeats itself, what does not yet cause damage but begins to deviate from what should be normal.

Prevention does not mean paranoia

There is, however, an important point to clarify. Taking weak signals seriously does not mean interpreting every anomaly as an imminent disaster. That would be the opposite mistake, equally unproductive. The value of weak signals lies not in alarmism, but in the ability to observe without trivialising.

The most useful prevention is not the one that turns everything into an emergency, but the one that builds contexts capable of noticing, recording, comparing and interpreting small signals before they combine into something more serious. In this sense, weak signals do not demand fear: they demand method.

In conclusion

Many serious problems do not begin with a major event, but with a small anomaly that nobody read as important. This is exactly what makes weak signals so relevant to safety: not because they offer absolute certainty, but because they force us to look at risk before it becomes obvious.

At work, in infrastructure, in public spaces and in domestic life, the difference between reacting late and preventing better often starts here: with the ability not to ignore what is still small simply because it is not yet making enough noise.