Artificial intelligence and machine learning techniques have a wide variety of uses that span virtually every discipline. Now, an international team of scientists has used these techniques to analyze millions of articles in newspapers around the world to provide forecasts of the risk of armed conflict.
Researchers have launched a website that offers a very up-to-date map to prevent outbreaks of violence up to a year before it occurs. The tool is therefore very useful for international organizations to manage the escalation of armed conflicts.
“We are hopeful that these forecasts can be used by those who must make public decisions and who are interested in reducing and preventing these cases that have not yet turned into a full-blown conflict,” the authors explain. In this way, it provides policy makers with an early warning before existing tensions escalate into armed conflict.
With the magnifying glass in low-risk countries
One of the strengths of the tool would be, according to its creators, that the method allows the extraction of subtle indicators in the news , even in countries at low risk and without a recent history of armed conflict. “The risk estimates provided here are ambitious because they emphasize these cases,” explain the scientists. Its forecasting model uses information from the Uppsala Conflict Data Program and newspaper articles published since 1989 to predict risks of violence.
Thus, they provide risk forecasts for two levels: the outbreak of any type of armed violence with fatalities, and the outbreak of armed conflict when a country exceeds the threshold of five fatalities per million inhabitants in one month.
The model has been extensively tested by scientists, as well as in a forecasting contest organized by the European Violence Early Warning System project with very positive results, according to the researchers.
Countries with very high risk of armed conflict
With the data updated today, the map shows that some of the countries in the world that are at increasing risk of an outbreak of armed violence (when it has not already broken out) are Mexico , Ukraine, Turkey, Bulgaria, Syria, Iraq, Iran, Armenia, Azerbaijan, Yemen, Afghanistan, Libya, Ethiopia, Sudan, Somalia, Democratic Republic of the Congo, Nigeria, Niger, Burkina Faso, Mali, and Myanmar .
The website has been financed within the Political Economy of Social Conflict: determinants and prevention project of the CSIC’s I-Link program. The general objective of this program is to stimulate, through its own resources, the collaboration of CSIC research groups with highly relevant international research groups. All of this within the scope of the Sustainable Development Goals, with the main objective of creating and consolidating consortia of international excellence to compete in future calls for Research and innovation (R&I) projects in the international arena.
