Mine safety has long been treated as a set of rules. Its next stage is the ability of people to recognize risk, manage it, and understand where that risk comes from.

Ivan Vereshchaka, PhD in Economics, adviser to Andromeda Demining Systems and author of DEMI, the world’s first AI chatbot for mine-safety education.

“I’ve Heard This Before” Does Not Mean “I Changed My Behavior”

A farmer leaves home at around six in the morning. The road is familiar. He has driven along it hundreds of times. Two kilometers from the village, the asphalt ends. Beyond it is a dirt road and a field that was not cultivated last season. There is a warning sign at the edge of the field. He sees it. He knows what it means. Last year, he even attended a mine-safety training session with around fifty fellow villagers at the local community center.

And yet he keeps driving.

This is not a story about someone who did not know the danger. It is a story about a decision made in half a second, one that previous training did nothing to influence.

Such decisions account for a significant share of civilian incidents in Ukraine. The most common circumstances include driving over explosive ordnance, working in fields, and attempting to remove discovered items independently. This is not exotic behavior, sabotage, or ignorance. It is everyday life.

Ukraine will have to live alongside explosive contamination for many years. Even after hostilities end, mines and unexploded ordnance will remain where people plow fields, build homes, go to school, collect firewood, and repair roads. An estimated six million people live in areas affected by this contamination.

Therefore, Explosive Ordnance Risk Education (EORE) cannot remain an emergency humanitarian communication campaign lasting only a few seasons.

“Do not approach. Do not touch. Report it” is necessary. For a country facing contamination on this scale, however, it is not enough.

The next generation of EORE must address a different question: how can we ensure that a person does not simply know about the danger, but can recognize risk, make the right decision, and perceive safe behavior as a normal part of life rather than an instruction?

And behind that lies an even more difficult question: can mine-safety education also build an understanding of the long-term cost of war?

What Cambodia Demonstrated: The Price of Metal Versus the Level of Knowledge

Countries affected by long-term contamination tend to follow the same trajectory. Initially, people are killed mainly through accidental contact. Later, deaths increasingly result from consciously accepted risk.

Cambodia is the best-documented example. People living in contaminated areas were well aware of the danger, yet they entered forests for resources, cultivated land, used dangerous routes, and collected scrap metal. Communities even developed their own informal survival rules, a form of local common sense shaped by decades of living alongside mines.

From the perspective of formal safety standards, some of these practices were dangerous. From the perspective of local residents, however, the issue was not ignorance but basic economic survival.

The most compelling evidence came from economics. A study by Roberts and Bilginsoy, published in World Development Perspectives, compared monthly scrap-metal prices at collection points near the Thai border with incident data. It found that the number of deliberate attempts to tamper with explosive ordnance responded almost one-to-one to changes in scrap-metal prices, while the number of accidental explosions did not depend on prices at all.

The authors reached a practical conclusion: to encourage villagers to report discovered items rather than dismantle them, it may be more effective to compensate them for the metal or allow them to collect scrap after the ordnance has been safely cleared.

This deserves emphasis. Behavior that had been targeted for decades through posters turned out to be a function of scrap-metal prices.

The lesson for Ukraine is direct. With each passing year, mine-action operators will increasingly deal not with people who do not know about the danger, but with people who know, understand, assess the risk, and nevertheless take it.

The question therefore needs to change. Not “How do we explain once again that this is dangerous?” but “Why does the person do it anyway?” and, most importantly, “What can actually change that decision?”

From Information Delivery to Risk Management

This is where EORE stops being merely communication.

If a farmer enters a dangerous field because there is no alternative land, another lecture will not solve the problem. If people collect scrap metal because they have no other source of income, a brochure does not eliminate the underlying cause. If a community uses a particular road because there is no safe alternative route, the appropriate solution may be anything except another training session.

Effective solutions may look completely different: warning signs, a safe route, prioritizing the survey of a particular area, alternative land, an economic support program, discussions with employers, public information about the actual status of a specific territory, or a decision by the local council.

This is no longer message delivery. It is risk management.

And the paradigm changes from the very first step.

Research the Decision Moment, Not Just the Audience

Traditional communication segments people in familiar categories: children, adolescents, farmers, internally displaced persons, and residents of de-occupied territories. This is necessary, but it is not enough.

It is much more important to understand the specific moment when a person makes a risky decision.

Let us return to the farmer from the beginning. He leaves home, reaches the point where the verified road ends, and stops. The decision takes place right there and lasts only a few seconds.

What is influencing him?

The economic pressure of the season. The experience of a neighbor who drove through the previous week and returned safely. The absence of an alternative field. The belief that “everyone drives here now.” Deadlines. An employer. Simple haste.

Analyzing a day in the life of an individual makes it possible to move from an information campaign toward designing safer behavior.

EORE must work not only with what people know, but with what they actually do during the critical second when a decision is made.

How People Actually Learn Safe Behavior

If the goal is behavioral change, different learning methods are required.

There is a qualitative difference between a person who has read a rule and a person who has practiced the correct response three times in a realistic scenario. Between these two stages lie several others: video, real-life stories, scenario analysis, and simulation.

This leads to a broader set of tools that EORE should use: scenario-based learning, storytelling, simulations, serious games, microlearning, VR and AR, adaptive learning, peer-to-peer education, and practical exercises.

The logic is the same as in fire safety, aviation, and industrial safety. When the danger is real, there is no time to recall a presentation. The required behavior must be learned to the point of becoming automatic.

Pilots do not simply read about engine failure. They practice it in a simulator until the response becomes automatic.

Digital formats have particular value for Ukraine: the country has a vast territory, a displaced population, and areas that are physically difficult to reach. The experience of the Danish Refugee Council with digital EORE courses demonstrates that scaling is possible.

There is another argument that is often overlooked: cost.

The pilot of the DEMI AI chatbot, developed by the author together with Andromeda Demining Systems, produced an increase in knowledge of more than 20% at a cost of approximately five cents per learner, according to the project’s internal assessment. The chatbot teaches through dialogue and adapts to the user’s age and context.

This changes the framework of the discussion. Adaptive learning no longer has to be an expensive experiment limited to individual pilot communities. It can become a tool deployed at national scale.

But digitalization is a tool, not the objective.

Online platforms are excellent for scaling knowledge. A live instructor is better at working with context: hearing the specific problems of a community, adapting the material, analyzing local scenarios, and understanding why people in that particular community take risks.

The optimal model for Ukraine is therefore likely to be hybrid: a digital platform combined with local instructors, practical scenarios, offline access, regular repetition, and feedback.

The Best Result Is an Event That Never Happened

Preventive education has a fundamental problem: its main result is invisible.

A person sees a suspicious object. Remembers the training. Does not approach it. Reports it.

And nothing happens.

There is no explosion, no victim, no entry in medical statistics, no news story.

That event that never happened is the principal result of EORE.

The same paradox is familiar in medicine, aviation, occupational safety, and disaster preparedness: the better prevention works, the less visible its results become.

This is why the number of training sessions conducted and people reached cannot remain the primary KPIs. These are activity indicators, not outcome indicators.

What needs to be measured is different: whether routes have changed, whether deliberate entry into dangerous areas has decreased, whether people report discoveries more quickly, whether risky practices have declined, and, most importantly, whether the effect remains six months after training.

This is an uncomfortable proposition for an industry accustomed to reporting reach. But without it, EORE will remain an activity that is easy to count and impossible to evaluate.

From a Safe Individual to a Safe Community

The next level is safety culture.

This is the point at which safe behavior stops being an individual decision and becomes a social norm.

In a mature safety culture, people report danger not because a poster tells them to, but because alternative behavior is already considered unacceptable by the community.

This is how aviation, nuclear, and industrial safety systems work: risks can be discussed, mistakes can be reported without punishment, dangerous behavior is not normalized, and safe behavior is reinforced by the group.

For Ukraine, this means that EORE must move beyond the work of mine-action operators and become integrated into education, community activities, agriculture, occupational safety, local planning, digital services, and the civil-protection system.

EORE must stop being a lesson delivered by a humanitarian organization.

It must become a permanent competence of society.

Children: 80% of Child Victims Are Adolescents

Children are usually viewed as those who need protection. The data present a more complicated picture.

According to UNICEF, adolescents aged 14–17 accounted for approximately 80% of child victims of mines and explosive remnants of war in Ukraine between February 2022 and March 2025, largely because they are more likely than younger children to deliberately explore dangerous locations. Adolescents from rural areas and low-income families face a higher risk.

Even among children, therefore, the problem is not simply lack of knowledge. It is behavior: curiosity, peer status, the desire to record a video, experiment, or prove oneself.

At the same time, children can become carriers of a new social norm.

The model is simple: child → school → family → community.

A child brings a rule home, asks parents uncomfortable questions, and draws attention to their behavior. Over time, what began as educational material can become a generational norm.

Mine-safety education in schools therefore should not be a one-time lesson. It should be part of broader education on how to recognize risk, make decisions under uncertainty, and take responsibility not only for oneself.

This is no longer simply mine awareness.

It is risk literacy.

The Mine as a Material Trace of War

We could stop here: teach people to live more safely, reduce incidents, and teach communities to manage risk.

But there is another level.

A mine is not simply a dangerous object. It is a material consequence of war.

A battle may last several hours. An operation may last several days. Its consequences remain in the ground for decades.

A mine can kill someone twenty years after the battle during which it was placed.

And the consequences extend far beyond those killed or injured. Contaminated land means lost agricultural production, blocked roads, delayed returns of displaced populations, devalued assets, inaccessible infrastructure, lost investment, and years of government and donor spending on survey and clearance.

EORE can therefore potentially explain not only “this is a dangerous object,” but also “this is what war looks like ten years after the fighting has ended.”

Can EORE Change Attitudes Toward War?

This is a difficult and largely unexplored question.

It must be approached carefully because EORE should not become political propaganda.

But there is a fundamental difference between propaganda and explaining cause-and-effect relationships.

There is no need to tell a child what they should think about war. Instead, they can be shown the facts: this object remained after the fighting; because of it, the community cannot use this land; clearance requires years and money; people continue to die after the fighting has ended.

The child can then see the chain for themselves:

war → contamination → risk → human losses → economic losses → development constraints.

This is education about the consequences of war.

And this is where EORE acquires a much broader social function.

In the public imagination, war exists through powerful images: military equipment, weapons, heroism, victories, and symbols. For children and adolescents, these images can be particularly powerful.

But war also has another dimension: a field that a farmer cannot enter for another ten years; a child who cannot play where their parents once played; a road that cannot be used; a community unable to return land to productive use; a person injured after the formal end of the war.

Understanding these consequences does not question the need for defense or a state’s right to defend itself. It simply adds what is often left out of the picture: war has a long-term cost, and civilians largely pay that cost after the fighting has ended.

Understanding this is not about valuing defense less. It is about understanding more clearly what reconstruction will cost and how many years it will take.

There is also a longer-term perspective that extends beyond mine safety.

Children who learn today how to recognize a dangerous object will make decisions twenty or thirty years from now in government, business, communities, and the military.

The question is what experience they will carry with them.

Will they remember war primarily as a collection of images and symbols, or will they have a practical understanding that every use of a weapon leaves consequences that civilians may have to deal with for decades?

The difference is not hypothetical.

Experiencing war does not automatically make a society either more cautious or more peaceful. A review of nearly two dozen studies covering more than 40 countries, published by Bauer and co-authors in the Journal of Economic Perspectives, identified a persistent pattern: exposure to violence can increase cohesion and willingness to cooperate within one’s own community, but this does not automatically translate into broader peace and can sometimes reinforce hostility toward outsiders.

In other words, trauma itself teaches nothing. What teaches is how society explains that trauma to the next generation.

That is why education about consequences is not merely awareness-raising. It is a long-term preventive mechanism.

Not to make the next generation less willing to defend itself, but to help it understand the cost of escalation more precisely where choices still exist.

Countries with long histories of contamination know this distinction from experience: memory that is not transformed into knowledge tends to fade after roughly one generation.

Five Generations of EORE: From Posters to Culture

Viewed over the long term, the development of the field can be divided into five levels.

EORE 1.0, awareness.
“Do not approach. Do not touch. Report.” Basic information about danger.

EORE 2.0, interactive learning.
Videos, digital platforms, simulations. The objective is better retention of knowledge.

EORE 3.0, behavioral risk education.
The focus is no longer the message, but behavior. Why does a person take a risk, and what could change their decision?

EORE 4.0, safety culture.
Safe behavior becomes a social norm, and the community itself participates in risk management.

EORE 5.0, consequence awareness.
A person understands not only how to avoid a mine, but why the mine is there and what consequences the war creates for decades.

The development logic is therefore:

recognizing danger → correct behavior → risk management → safety culture → awareness of the consequences of war.

What This Means for Ukraine

Ukraine has a rare opportunity not simply to scale conventional EORE, but to create its next generation.

This requires changing the way the task is defined.

Not “How many people did we train?” but “How did their behavior change?”

Not “How many lessons were conducted in schools?” but “Has a lasting model of safe behavior been established among children?”

Not “Does the farmer know that the field may be dangerous?” but “Why does the farmer still enter it, and what could create a safe alternative?”

Not “How do we avoid explosive ordnance?” but “How do we build a society capable of systematically managing residual risk for decades?”

And finally: can education about the concrete consequences of war help raise a generation that better understands its true cost?

In mine action, it is easy to count cleared hectares, removed items, and people who have received training.

It is much harder to count the person who did not take the dangerous step.

The child who did not pick up the object.

The farmer who turned around at the edge of the field.

The family that changed its route.

The community that eliminated a risk before it became a tragedy.

The best result of preventive education is an event that never happened.

But perhaps there is a higher-order result.

When education changes not merely an individual action, but a way of thinking. When people approach risk, responsibility, and the safety of others differently, and ultimately understand why that risk exists in the first place.

The future of EORE in Ukraine is therefore not about making millions of people memorize the phrase “do not approach, do not touch” one more time.

The task is broader: to teach society to recognize risk, manage it, resist becoming accustomed to it, and understand the long-term cost of the war that created that risk.

Disclosure of Interest

The author of this article is the author and developer of the DEMI chatbot (Andromeda Demining Systems) mentioned in the text. The pilot data are presented based on the project’s internal assessment.

Sources

• Landmine and Cluster Munition Monitor, Ukraine profile: estimate of approximately six million people living in areas affected by contamination.
• ACAPS, thematic report on the humanitarian consequences of mine contamination in Ukraine.
• UNICEF, as cited by Just Security: adolescents aged 14–17 accounted for approximately 80% of child victims between February 2022 and March 2025.
• Roberts, W., Bilginsoy, C. “Risking life and limb in the global economy: Scrap metal price and landmine/UXO incidents in Cambodia.” World Development Perspectives, 2016.
• Bauer, M., Blattman, C., Chytilová, J., Henrich, J., Miguel, E., Mitts, T. “Can War Foster Cooperation?” Journal of Economic Perspectives, 2016.