Few civil service in-trays are as busy right now as those at the Ministry of Defence. Since the end of the Second World War, peace in Europe has dominated industrial, military and government thought patterns – now the UK and its Western and NATO partners face multiple geopolitical challenges. These challenges have prompted an internal re-assessment of military readiness, from administration to ammunition, from procurement to front line. Amid this realignment, one thing is clear: artificial intelligence (AI) will play a fundamental role in how the MoD functions. IBM cites General Sir Roly Walker KCB DSO, chief of the general staff, who has said that the British Army must exploit advanced
technologies, such as AI and autonomous systems, as it prepares for the ‘unfair fight’ against bad state and other actors.
The game-changing nature of AI in an ever evolving, unpredictable world was the focus of a senior defence leaders’ forum hosted by IBM, in partnership with Civil Service World, attended by top-tier civil servants. IBM has a longstanding relationship with the MoD, running Guardian – the command-and-control system that helps the RAF defend UK and NATO airspace, and last year, IBM signed a £30m contract with the MoD for various civilian and commercial IT systems. According to Pat McFadden MP,
the Chancellor of the Duchy of Lancaster, the UK government is planning the biggest reform of the department in over 50 years. Last November, the Minister for Defence Procurement and Industry, Maria Eagle, announced that the MoD is developing a ‘productivity portfolio’ to explore the potential of emerging technologies, including AI-driven tools and generative AI to automate and accelerate routine tasks, from operations and policy work through to defence-specific activities driven by data
to enhance the speed of decision-making, optimise logistics and increase the availability of military capabilities.
Paul Macpherson
Senior Partner and VP of the UK
Defence team at IBM Consulting
Despite its potential, AI adoption in the MoD is not without challenges. To Paul Macpherson, Senior Partner and VP of the UK Defence team at IBM Consulting, there is an understandable concern regarding whether decision makers can really trust the technology to make critical decisions, and this could be a significant blocker to MoD having a more widespread and consistent application of AI.
Innovating at pace
A recent IBM report, AI Decision Advantage for Defence, found that while defence leaders recognised that AI would improve their operational capability and readiness, it also concluded that “AI skill shortages, data governance issues, and ethical challenges continue to be speed bumps.”.
The MoD is a huge department, employing more than 248,000 people in the armed forces and civilians combined. IBM believes that for AI to be a catalyst for rapid change, a departmental mindset shift is key. “We need to industrialise AI across the MoD from end to end,” says Col Chambers, IBM UK Defence Managing Director for Technology. “If you want to develop a modern radar system to
detect hypersonic missiles, you can’t wait 10 years for proof of concept. You need it ASAP and you need AI to get you there.”
The UK’s Ajax armoured vehicle programme took 14 years to reach field trials, whereas, in extremis, Ukraine developed and deployed the Inguar-3 multi-role vehicle within two years. “The MoD was seeking perfection because it had the luxury of time,” Chambers says. “Now, it needs to innovate at pace and with nerve. We have always innovated quickly in wartime; radar was a great example of that in WW2 and we need to have that same mindset now – and we have the tools to do that.”
The first crucial steps
For now, the immediate AI’s potential is in transforming back-office functions, from information management to inventory management, HR, legal and finance. “The government wants departmental leaders to be accountable for efficiency savings,” says Macpherson. “In our experience with AI in large organisations, savings can reach 35%-40%.”
Col Chambers
IBM UK Defence Managing
Director for technology
A key issue for the MoD is its reliance on legacy systems that operate in silos, and Chambers says IBM, in collaboration with MoD, can help identify and address key challenges. “This structure limits integration and interoperability,” he explains. “AI requires access to aggregated data from across these
silos, enabling a comprehensive synergy and richer strategic view. To prevent additional complexity and duplication, data must remain within its existing structures. This approach effectively provides a multi-dimensional perspective on the data, enhancing efficiency and operational effectiveness.”
IBM brings expertise from sectors beyond defence, such as banking and oil and gas, where AI has transformed operations. “There are a lot of shiny AI tools, but strong governance and data management are fundamental,” says Macpherson. “You need confidence that AI is telling you the right decision to make. If you put rubbish in the system, you get rubbish out.”
AI hallucinations – where models generate misleading results – are a known risk. However, over the years, IBM has implemented robust governance frameworks to mitigate such issues. “AI is mature now,” says Macpherson. “We know what the bear traps are, and that if governance is in place, then you can trust the data.” He notes that small language models, tailored for specific defence applications, can
be more effective than large-scale general models. “The real advantage of AI does not lie in the technology itself, but in how it is deployed and made accessible to users to maximise its benefits,” Chambers adds.
Another challenge to AI adoption in the MoD is its hierarchical structure, which makes it difficult to implement department-wide strategies. “People often call for change at the end of their careers when they have nothing to lose,” says Chambers. “We need to empower internal innovators, those who understand the value of change and can drive transformation.”
The future of AI in defence
AI also presents workforce implications. Some civil servants fear automation could lead to redundancies, but Chambers argues that AI will instead alleviate workforce shortages. “The MoD struggles with recruitment and retention,” he says. “People sign up for specialised trades but end up doing mundane tasks, leading to attrition. AI can free them from these tasks, allowing them to upskill and find more fulfilling roles.”
Concerns about ethical AI use in defence also persist. However, Chambers dismisses the notion that civil servants face insurmountable ethical dilemmas. “Mistakes will inevitably occur – some data may slip through the net – but under the current system, human decision-making introduces even more
errors, often with inherent biases. That is simply human nature.”
Chambers emphasises that AI governance and ethics tools are already available to ensure a transparent, organisation-wide view of data, models, and decision-making processes. “Civil servants operate within a highly regulated framework; they understand the rules and must apply them while adopting AI in a structured, incremental manner. The best approach is to start with back-office functions, identify productivity gains in low-risk areas, and then scale up to more complex use cases once confidence is established.”
Today, the UK government remains in an exploratory phase with AI. “There is a lot of experimentation, but AI needs to be fully industrialised across the MoD to deliver the ‘unfair advantage’ General Walker seeks,” says Chambers. “The UK is far from where it wants to be, but with the right approach, progress
can be significantly accelerated.”
Sign up for IBM’s Senior Defence Leaders Forum to hear from MoD's David Blackall, DESNZ and DSIT CDIO Karl Hoods, and other leaders from across defence and Nato sharing provocative ideas and real-life examples digital transformation in defence. Event takes place on 3rd March.