Why it matters

An internet search using the keyword ‘Artificial Intelligence’ (AI) yields close to 95 million results, of which some 3 million only on the term’s definition, which gives an idea of the growing popularity of the term.

AI can be understood as the theoretical creation and development of computer systems or algorithms able to perform tasks normally requiring human intelligence. There is a continuous, controversial debate going on in the media: on the one hand, about the potential of artificial Intelligence as a game changer that will bring huge benefits to humankind; and, on the other, on the supposed threat it poses to our civilization given that the impact of AI in the future is difficult to evaluate at this time. 

There are already calls for ethical regulation before we lose control on this technology. Commercial companies making huge profits on the global market are driving innovation in this field and developing new algorithms to provide intelligence through different applications exceeding recognition of images, voice and text.

In the last century, the arrival of digital computers made it possible to perform mathematical operations and data storage at a rate far beyond the capabilities of human beings. It then became possible to develop very complex algorithms and eventually programme machines in a way that allows them to learn and provide solutions comparable to what we call intelligence in human beings. This gave computer based systems the possibility to learn from data, the environment and from their own errors; something known as ‘cognitive computing’. 

Deep Learning, a technology based on specific kinds of Neural Networks, is responsible for the quantum leap observed in the field since 2009. It uses data analysis to predict trends and discover hidden information and patterns in the ocean of data provided by existing networks and sensors. The limits of this technology are not currently known but there are already some warning voices out there urging society to be prepared for an apocalypse as a result of this technology in a not so distant future.  

What the EDA does

The EDA’s network of experts working in the ‘Radar’ Capability Technology Group (CAPTECH) first drew attention to the growing importance of Deep Learning (DL) technologies during a workshop held in 2015, after which the ‘DEEPLEARN’ project was launched.

The aim was to understand the possibilities for applying these algorithms in the Defence domain, motivated by the very successful results obtained in the civil area by companies like Google, Apple or Facebook. In fact, the use of massive amounts of data seemed to be needed for the correct training and functioning of DL tools and the project aimed to show the operational limits of creating a mathematical framework to understand the applications in defence. Different applications like the detection of malicious traffic in encrypted networks for cyberdefence, the identification of gaits and gestures from people walking, or the detection of anomalous behaviour in maritime vessels traffic are currently being analysed in the realm of this project. 

The way ahead

Another AI workshop in the field of Communication and Information Systems (CIS) together with Modelling and Simulation (M&S) was held in May 2017 aiming to find future areas of collaboration where AI can play a significant role. 

Member States’ participation in collaborative projects is key to developing a European capability based on AI, for example in areas such as  support to decision making for the Commanders, extraction of relevant semantic information to support Military Intelligence, creation of realistic behaviour for opposite computer generated forces or intelligence for autonomous operation of unmanned vehicles. 

The potential areas of application are so wide that joint investment programmes should be established to allow all Member States’ Forces to have access to these new capabilities at all possible levels of operation.

Previous article

A constant eye on the future: identifying Europe’s capability requirements for 2035

Next article

Defence Internet of Things (dIoT)