Your browser version is outdated. We recommend that you update your browser to the latest version.

SIGN UP to our free eNewsletter

Fields marked with * are required.

DEFSECmedia

DEFSEC Media is New Zealand's defence, security and fire B2B/B2G publishing group. Our leading magazines, Line of DefenceNZ Security and Fire NZ are read by key business, government and military decision makers and influencers. This website is the online home of cutting-edge content from each of our titles.

Itron

Innovation Central

Automation & Electronics

AI and machine learning: A new kind of military intelligence

Line of Defence Magazine, Summer 2017/18

AI's defence applications are not all about Terminator robots... in fact quite the contrary.AI's defence applications are not all about Terminator robots... in fact quite the contrary.

 

Evan Butler-Jones, Director, Defence Product Line, Aviation & Defence Business Unit for IFS, writes that although the defence applications of artificial intelligence is a hot – and often hyped – topic, AI is more about enabling safe and efficient missions than ‘Terminator’ robots.

 

Artificial intelligence (AI) and machine learning is one of the hottest technology topics of the moment. It has been grabbing headlines and airtime throughout 2017 – and it looks set to continue into the next year.

Gartner has urged “the risk of businesses becoming non-competitive or even obsolete by ignoring AI is high” and according to IDC research the AI market is set to be worth $46bn by 2020. This disruption will be felt in the defence industry too.

This discussion always raises the spectre of autonomous weapons, a topic with many ethical implications that deserves more formal attention.  However, there are other less-often considered uses of AI that nonetheless have the potential to dramatically impact global defence activities.

 

No longer droning on – Empowering autonomous equipment

Drones are already taking hold in the defence environment. Efficiency savings can be made by making use of equipment which is cheaper, more versatile and doesn’t involve sending soldiers onto the frontline. In late 2016, the British Royal Navy launched a full-scale exercise made up entirely of unmanned equipment — nine different assets were flying, driving and sailing while interacting with each other.

This equipment was remote controlled, but AI can take drone technology to the next level, helping drones navigate and make sense of the world they operate in.

A start-up called Exyn Technologies Inc. recently revealed AI software that enables drones to fly autonomously, even in dark, obstacle-filled environments or beyond the reaches of GPS. A spin-out of the University of Pennsylvania’s GRASP Labs, Exyn uses sensor fusion to give drones situational awareness much like a human.

Couple this with image recognition technology and there is potential for serious improvements in efficiency and mission success. Surveillance drones for example, produce a huge amount of data on every mission, which must be pored over and deciphered manually.

However, with image recognition the drones themselves can provide detailed contextual analysis in real-time, leaving mission control to concentrate on strategy. Far from being a future consideration, this is already being explored in transport and civil infrastructure, using Neural Networks to detect and classify objects in coastline observation, fire detection, and three-dimensional mapping.

 

Enjoying this article? Consider a subscription to the print edition of Line of Defence.

 

Bolstering cybersecurity 

AI has close links to one of the hottest subjects circling the defence industry right now – cybersecurity. Finding malicious activity in the streams of data moving in and out of defence networks is vital but can be similar to finding a needle in a haystack. One report found that it can take an average of six months for organisations to detect a data breach. 

Enter the fast-developing power of AI. Machine-learning algorithms are being used more and more to detect and block security breaches. Researchers at the MIT Computer Science and Artificial Intelligence Lab (CSAIL) developed an AI system that uses supervised learning to find and report security threats.

Over the space of a few weeks, analysts trained the system to distinguish the difference between real security threats and false alarms. After a while, the system learned to reduce false positives by flagging significant events that analysts could then investigate.

In a world where cyber-attacks and data breaches are becoming an almost daily occurrence, AI could provide a safe haven for defence organisations.

 

A worthy adversary and ally – Revolutionising military training 

One of the clear advantages of artificial intelligence is its potential for constant improvement. This is particularly true in training environments where static systems cannot evolve enough to consistently challenge military personnel.

In a simulation in 2016 an AI pilot shot down a decorated former US Air Force pilot with over a decade of in-field experience. It was powerful enough to beat all other AI pilots and shoot down the US pilot in every simulated ‘dogfight’. The system approaches complex problems much like a human would. Larger tasks are broken into smaller subtasks, which include high-level tactics, firing, evasion, and defensiveness. It can make complex decisions with extreme speed by only considering the variables it considers most relevant. 

The AI program has since been re-commissioned to join forces with USAF pilots. It can now be used as a friendly co-pilot system to help human pilots using the simulator with the aim of improving reaction and decision-making time to implement more effective military actions.

 

Intelligent decisions, from procurement to the frontline

The complexity and mission criticality of the defence supply chain cannot be underestimated. Under-planning and under-stocking can be the difference between success and failure in a military context, obtaining a replacement part could take months, when it's often seconds that define military success and failure. 

Intelligent systems that can parse through the massive amounts of logistics and maintenance data held by defence contractors can identify optimisation opportunities. AI can be coupled with current connectivity innovations in defence equipment, think Health Usage Monitoring Systems and the proliferation of data being fed into supply chains by IoT enabled sensors.

These analytics are leading the development of predictive capabilities. New models produced by automated machine learning will take all the data from the smart infrastructure to produce alerts about failing assets or anomalous readings ahead of asset malfunction, with detailed information about what corrective actions need to be taken before operations are disrupted by component failure.

 

Working together, not exterminating

Much of the hype around AI in the military centres around Star-Wars like robots and droid ‘super-machines’. In reality, application of AI in defence is far more about enabling missions to be executed safely, effectively and efficiently. 

 

Back to Defence

Share on Social Media

Follow us...

We recommend on YouTube...