How AI can improve military decision support

A new doctoral thesis from the Swedish Defence University demonstrates how artificial intelligence (AI) can enhance military decision-making in complex situations and contribute to increased precision and efficiency.

By developing and testing two AI models, Lieutenant Colonel Peter Bovet Emanuel, PhD candidate in War Studies at the Swedish Defence University, shows how AI can support military decision-makers in challenging and time-critical scenarios.

“I’m exploring how AI can be integrated into the decision-making process to increase speed, accuracy, and adaptability", he says.

Two models for military decision support

The first model helps optimise intelligence gathering against high-priority targets such as radar systems. It simplifies the process of collecting the right information—faster and more accurately. The model can also predict where an adversary’s weapon systems might be located. This allows for a more efficient allocation of tasks to sensor systems and provides decision-makers with a stronger basis for informed decisions.

“This model uses neural networks and cutting-edge AI technology, deep learning, which adapts and learns as more data becomes available”, says Peter Bovet Emanuel.

The model has been trained on large volumes of satellite imagery and has learned how topographical information influences the propagation of radio signals. It involves multi-step problem-solving where the neural network helps recommend the most suitable locations for radar systems.

Decision support for joint targeting

The second model employs a more traditional AI approach, using optimisation algorithms to help select appropriate targets, even when there are many options and resources are limited. It can be adjusted to reflect a commander’s priorities and supports the entire decision-making process.

“This is about decision support within the process NATO refers to as ‘joint targeting’. It’s a fundamental process at operational and tactical command levels focused on identifying and prioritising targets, and matching them with suitable means of engagement to achieve the desired effects. It’s a thorough analytical method aimed at optimising the use of one’s own capabilities”, explains Peter Bovet Emanuel.

Both models are based on the OODA loop (Observe, Orient, Decide, Act), a framework used to describe and understand decision-making in rapidly changing environments.

“AI can offer military decision-makers a broader range of options and create opportunities to take the initiative more quickly and adapt to changing conditions”, he adds.

AI’s impact on military decision-making

The study is of importance to both practitioners and researchers. For military decision-makers, it demonstrates how AI can be applied practically in current and future targeting work. For researchers, the thesis contributes new knowledge about how AI affects decision-making, command, and coordination in military contexts.

“My research shows that AI has the potential to change the nature of decision-making—from being something only humans do, to becoming a collaboration between human and machine. Armed forces must therefore plan for how AI can become a part of future warfare.”

He also stresses the need for further research into the ethical implications and how command authority is affected in AI-supported systems.

“The dominant view today is that humans remain in control and make the final decisions. My research challenges that assumption and may have significant implications for how command and leadership are structured in the future.”