Challenge of OSINT in Developing Countries 

This post is also available in: עברית (Hebrew)

Valuable information can be pulled from media reports, public financial information and social media posts. Websites track user activity, and smartphones are constantly gobbling up information about their users, from geolocations to search histories and more. These serve as open-source intelligence, known as OSINT. Now, artificial intelligence tools will help analysts make sense of this torrent of publicly available data and turn it into usable OSINT.

However, in developing countries, there is a less comprehensive and representative picture of what the population looks like through the same types of techniques that are being used in the U.S.

The U.S. military and intelligence community are increasingly interested in leveraging OSINT for predictive analysis, helping warn regional commanders of upcoming political protests, political violence, extremist attacks or other kinds of security-related events could take place, Notably, the Army awarded BAE Systems a $437 million task order for open source intelligence support in October, as reported by c4isrnet.com.

FRAYM, a geospatial data and analytics company, wants to create data-rich analysis in data-poor areas in order to create usable and reliable OSINT.

According to their CEO, Ben Leo, “What we do is we gobble up the very high quality, underutilized datasets that are out there. We bring in additional public datasets and we bring them all together using our AI/ML algorithms to produce this hyper-local data at scale.”

The company takes geotagged household data and feeds that into its machine learning algorithm, and from there it can then produce data down to a 1 km x 1 km grid-level across dozens of characteristics, such as religion, ethnicity, language, age, education access, electricity, media consumption and more.

In the past, there were only two ways to make predictions in data-poor areas. First, analysts could monitor events through social media. While that can help commanders understand the situation on the ground, it has very limited predictive power. The other method was to basically plot out how and where events have unfolded in the past and try to find correlations that can predict future events. That, too, is very limited.

But with access to proprietary data the U.S. government does not have, Leo says his company has been able to create a unique source of OSINT for data starved areas.