The perils of open source data
Military agencies are in an on-going battle to maximise the benefits of commercial open source data while avoiding the potentially devastating intelligence pitfalls.
The security risks associated with open source data were starkly highlighted by the release of Strava’s Global Heatmap.
The map, which shows the routes travelled by users of its exercise tracking product, inadvertently exposed sensitive information about American and allied military bases and troop movements in conflict zones across the world.
The map highlights a number of well-known military bases across Iraq and Syria, where Western soldiers have been stationed as part of Operation Inherent Resolve.
However, it is the movements of troops outside their bases, their patrol and supply routes and smaller camps not previously known about which could offer valuable information to enemy forces.
At the recent Defence Geospatial Intelligence conference in London, military and industry leaders discussed how security challenges can be overcome to enable better exploitation of the vast reams of commercial data available to military and intelligence agencies.
However, as Maj Gen James Hockenhull, director of cyber intelligence and information integration at the UK MoD noted, the relationship between the military and industry requires improvement.
With incidents such as the Strava heat map, military users of commercial geospatial systems remain sceptical about the security and reliability of the data being collected and disseminated.
However, the proliferation of commercial satellites offers a huge potential for militaries to access near real-time, high-resolution imagery within government spending constraints.
Further developments in artificial intelligence, machine learning and deep learning algorithms have led to a significant leap in the efficiency, and outsourcing, of geospatial intelligence analysis to companies willing to invest heavily in the technologies.
These include companies such as Esri and DigitalGlobe who are developing deep learning algorithms to enable automated identification of a wide range of objects.
This could provide intelligence agencies with rapid and accurate strategic information, such as the movements of enemy military equipment, troops, or weapons testing locations, such as the site of the North Korean missile test pictured above.