Using satellite, drone and other data feeds to monitor potential risk factors on a macro scale across a potential battlespace. This might be a country - North Korea - and it might involve specific issues - in this case deployment of nuclear missiles. The potential scale, and opportunities for spoofing and mistakes, is huge.

Radar is used to scan individuals to identify shape and metallic content of devices being carried, including those concealed under clothing. Machine learning is then applied to the data to identify potential weapons - be they guns, knives or even bombs. Inevitably there is, at this stage, a high risk of false positives that will potentially impact on human responses to triggered alerts.