AI in Defence: Advantage or Dependency?
Artificial Intelligence (AI) is reshaping the character of modern warfare. No longer confined to logistics or surveillance, AI is influential across operation planning, intelligence analysis and even frontline decision-making. Recent developments, ranging from UK-backed initiatives in Ukraine to new battlefield medical trials, demonstrate the pace of innovation and the growing strategic importance of AI. For governments and defence organisations, the question it not whether to adopt AI, but how it can be effectively used to enhance resilience and long-term security.
The war in Ukraine demonstrates how AI is transforming warfare. It is critical in drone warfare, where it can help drones identify and track targets – even when communications are jammed. According to Reuters*, AI-operated FPV drones could raise the strike rate from 30-50% to over 80%. AI is also used to handle enormous volumes of battlefield data, include satellite imagery, drone footage and ground reports, as well as open sources like social media which can help locate enemy forces. Its importance is increasingly recognised as central to the conflict, reflected in the UK’s £500,000 investment in a Centre of Excellence within the Ukrainian Ministry of Defence. The aim of this is to ensure AI is being used to its full potential in the conflict, and the UK benefit from the lessons learned for our own defence.
Closer to home, a recent military medic AI trial provides a tangible example of how AI is moving from theory to operational reality. The trial, where DSTL collaborated with the US Defense Advanced Research Projects Agency (DARPA), is a part of a long-term investigation into whether AI is trusted by medics in high pressure situations, and whether it could help save lives. While post-trial analysis is ongoing, the concept highlights the possibilities of human-machine collaboration. Rather than replacing human judgement, AI can be used to help medics make faster and more consistent decisions under extreme conditions. If proved effective, increased trust in AI could be transformative for battlefield medicine.
Nevertheless, the use of AI tools in high stakes environments like warfare come with serious risks. AI models can be trained on inaccurate data, and small errors can lead to deadly consequences. As statistical models, they won’t always get be correct. This underlines the importance of ensuring that models are sufficiently mature before deployment. For AI to be used effectively and ethically, there needs to be robust governance, clear guidelines and human oversight to prevent potentially fatal errors.
As AI becomes embedded in defence operations, the UK faces a strategic choice. While collaboration with allies and industry partners across the world remains essential, over-reliance on external providers – particularly large technology firms – risks creating dependencies in an already unstable world. The UK should prioritise the development of sovereign AI capabilities. This means investing in advanced technologies, but also domestic infrastructure, talent and industrial capacity. Crucially, it requires ensuring defence AI systems, and the data that powers them, remain under national control. We should not replace reliance on international partners with reliance on global technology providers. After all, strategic autonomy is not about isolation, but about retaining the ability to act independently when it matters most.
*Reuters, Ukraine rushes to create AI enabled war drones, https://www.reuters.com/technology/artificial-intelligence/ukraine-rushes-create-ai-enabled-war-drones-2024-07-18/