How can military training be described?

Artificial intelligenceHow AI systems are changing military training

Georg Klein (rehearsed): "My responsibility is training. That is why I look very closely at this area. There are already many areas of application for artificial intelligence, including in the military sector. I am particularly interested in how I can use this to prepare people for their jobs in peace, crisis and war . "

Manfred Kloiber: Georg Klein, Brigadier General of the Bundeswehr, justified his interest in artificial intelligence methods. As a department head at the Armed Forces Base Command, Klein is responsible for training the soldiers. And that is exactly what it was all about on Tuesday and Wednesday of this week in Bonn at the conference of the German Society for Defense Technology, about artificial intelligence in training, its opportunities and risks. Peter Welchering, you attended the conference. How do the soldiers and defense experts discuss AI in training in the Bundeswehr?

Peter Welchering: When it comes to the use of artificial intelligence methods in the Bundeswehr, we are talking about three areas: In operational use, it is about unmanned systems, autonomous weapon systems and robots. In the area of ​​reconnaissance, we are dealing with predictions about the behavior of opposing armies, but also with recognizing tactical requirements in action. And soldiers need to be trained for these two areas. The primary interest in artificial intelligence lies precisely in the training; this is training that conveys the skills and abilities to be able to deal with the AI ​​systems used. Then there is the use of AI in learning management, i.e. suggesting suitable training courses for the individual soldier and adapting them accordingly for individual learning. AI methods are also intended to ensure that the learning soldier is not overwhelmed, but also not under-challenged, when learning. This also includes the use of AI methods to answer the question: which post is suitable for the recruit? AI as an extension of personnel management, so to speak at the interface between learning management and personnel management. And another area of ​​its own is the use of AI-supported computer games.

Two digitized divisions by 2027

Kloiber: So the military are actually interested in exactly the same applications as companies or governments. And they also have to struggle with the same problems when setting up and introducing AI systems. This was also discussed intensively at the Bonn conference on artificial intelligence in training. How strong are the, I would say, business processes of the Bundeswehr have already been digitized here and what role do AI methods play in this, Peter Welchering?

Which ring: There is experience with unmanned systems among the mountain fighters. There are first attempts to use operational data from weapon systems for predictive maintenance. There are analysis systems for the education. At the end of this year, i.e. 2019, the concept "AI in the Army" should finally be worked out. And in 2027 there should be two so-called digitized divisions. The digital command post, which we have already reported on in Computer and Communication, has so far not been seen as a model for success. In this respect, it made me sit up and take notice that the subtitle for the Bonn conference was 'Opportunities and Risks for the Armed Forces'. In fact, the use of AI systems in the Bundeswehr carries a lot of risks. And these should be discussed and clarified beforehand now that the equipment with AI systems is imminent.

Defense against autonomous weapon systems

Kloiber: Which risks and problems do the general staff want to tackle first?

Which ring: The very fundamental problem facing the Bundeswehr is that of many companies. First of all, the very simple question must be answered: What data is needed for what? So there is still no data concept. But that is a prerequisite if we want to talk about the use of AI methods. Then it must be clarified how and on which AI systems the soldiers should be trained. And the operational principles and the available weapon systems must be adequately described and taken into account. Because that's what the training is based on.

Kloiber: But does that also mean that AI methods for autonomous weapon systems are also an issue here?

Which ring: Yes, of course, there are autonomous weapon systems that you have to deal with. Concepts are needed for defense against autonomous weapon systems. And for these concepts, in turn, the methods of machine learning and pattern recognition must be known and learned. And this knowledge has to be brought into the Bundeswehr under considerable time pressure. And there you can already see today: AI methods not only require an adaptation of the management processes, but they will massively change everyday life in the Bundeswehr.

Kloiber: I spoke to Peter Welchering about the use of artificial intelligence in the Bundeswehr, thank you!