The Pentagon is experimenting with artificial intelligence that can predict events "days ahead"

The Pentagon is experimenting with artificial intelligence that can predict events "days ahead"
The Pentagon is experimenting with artificial intelligence that can predict events "days ahead"
Anonim

If you're wondering how advanced artificial intelligence (AI) systems are becoming, then be aware that the US military is testing an experimental AI network that aims to identify probable future events that deserve close attention, and several days before they occur.

The series of tests is called Global Information Dominance Experiments (GIDE), and they combine data from a wide variety of sources, including satellite imagery, intelligence reports, field sensors, radars, and more.

Cloud computing also plays an important role in this system, ensuring the efficient processing of huge amounts of data collected from around the world, and the subsequent access to them by those military officials and departments who need them.

"GIDE, experiments in global information dominance, embody fundamental changes in the way we use information and data to expand the decision-making space for leaders from tactical to strategic levels - not only for military leaders, but also empowering our civilian leaders." - explained US Air Force General Glen D. VanHerk at a press briefing last week.

The idea is to anticipate other countries' actions in advance, which means that deterrents and precautions can be taken before hostilities start or before hostilities have a chance to flare up.

In fact, the achievements obtained as a result of the work of AI are not as great as it might seem - for example, if there is a preparation for the exit of a submarine from the port, then it is quite obvious that it is heading to sea. Where AI really helps is in using machine learning to identify and correlate all of this information much faster than humans can.

Another example is the number of cars in a car park, perhaps at a military base or research station. If the AI ​​notices increased activity, it can report it to other parts of the system, where this information is then analyzed as part of a huge data set.

"The data exists," VanHerk said. "We make this data available, so that this data is transferred to the cloud, where machine learning and artificial intelligence examine it. They process it very quickly and provide decision makers with what I call decision excellence."

"This gives us several days of early warning and the ability to respond. Where in the past we might not have been able to look with an analyst at a GEOINT satellite imagery, we now do it in minutes or near real time."

Understandably, the US doesn't talk too much about exactly how these new AI systems work or how they process the information they collect, but the end result is more data processed in a shorter time. The third GIDE trial was recently completed and a fourth is planned.

While the experiments look a bit like Minority Report, where people are arrested for crimes before they are committed, at the moment, the experiments are viewed by officials as a form of information gathering with increased efficiency rather than a way of looking into the future.

VanHerk emphasizes that people still make all decisions based on the data that machine learning systems produce, and says that the AI ​​being developed is likely to ultimately contribute to de-escalation, and not vice versa.

“The ability to see days ahead creates space for decision making,” VanHerk said. "For me, as the operational commander, there is a decision space to potentially deploy forces to create deterrence options and submit that to the Chiefs of Staff or even the President."

Popular by topic