Government response to Developing AI capacity and expertise in UK Defence

Third Special Report

The Defence Committee published its Second Report of Session 2024-25, Developing AI capacity and expertise in UK defence (HC 590), on 10 January 2025. The Government Response was received on 24 March 2025 and is appended below.

Appendix: Government Response

Introduction

1. The Government welcomes the report by the House of Commons Defence Committee: ‘Developing AI capacity and expertise in UK defence’, published 10 January 2025 (HC 590), and is grateful to all who gave evidence in the preparation of the report. We welcome the Committee’s thorough and thought-provoking analysis. These issues will be of increasing importance in the years to come as AI-enabled military systems and capabilities become more common.

2. AI, and the autonomy it enables, has the potential to transform every aspect of Defence. It is becoming a critical enabler for a modern, relevant military force and is already changing the character of warfare. Allies and adversaries are investing heavily in AI as a force multiplier and a wide range of early-stage applications are already being deployed on battlefields (e.g. in Ukraine). The credibility of future UK military capabilities, and our ability to deter or confront threats, will increasingly depend on our ability to develop, integrate, and scale AI-enabled solutions at pace across the full breadth of the Defence enterprise.

3. The MOD has already taken steps to harness AI technologies, and in many ways compares favourably with peers. This includes a Defence AI Strategy (2022) and Defence AI Centre (DAIC) which was established in 2022. This Government is determined to accelerate the use of AI across defence and there are currently over 400 AI-centric projects underway across the department.

4. We are widely respected for our approach to Responsible AI and as an acknowledged thought-leader in related international dialogues; the MOD are pursuing AI-enabled capabilities in both the frontline battlespace and the corporate business space. AI is already being used to support operations, enhance intelligence analysis, improve equipment availability, and pursue productivity savings. Specific examples include:

• Typhoon Predictive Maintenance Optimisation – this tool uses GenAI to assign meaning to narrative data in millions of historic maintenance records. This data is then analysed to provide recommendations to optimize future maintenance schedules and avoid unnecessary inspections. Recommendations that have been validated and approved to date are saving c. 4750 maintenance hours annually and increasing platform availability.
• SPOTTER – exploits computer vision techniques to improve the prioritisation of scarce analyst capacity to analyse overheard imagery from satellite.
• AI for Acoustics – processes large volumes of underwater acoustic data at pace for operational effect, as part of AUKUS collaboration with Australian and US partners, UK industry, and Dstl.
• Project IKENA – is an incubator for solving Digital Transformation challenges associated with the Permanent Joint Headquarters. Embedded AI experts and user researchers in the Transformation Team have been critical in evolving key business processes to leverage AI for more efficient and effective outcomes.

5. The MOD recognises the important role the UK’s Defence AI sector needs to play in promoting growth. The Department’s AI adoption is situated in the context of the recent AI Opportunities Action Plan published by the Department for Science, Innovation and Technology (DSIT), which provides a roadmap for securing UK leadership in AI. This includes national initiatives to enhance AI skills and attract talent, increase UK super-compute capacity, establish a UK Data Library, create AI Growth Zones, and generate world-class capabilities in advanced AI. The MOD will benefit from national uplifts in AI skills, infrastructure and industrial capacity and will actively work alongside other government departments and industry experts to share innovation and best practice.

6. The Committee’s recommendations provide a timely challenge to inform the MOD’s approach to the prioritisation of key activities, including AI skills development, ways of working with industry, and putting technical and policy enablers in place to accelerate our responsible AI adoption. Whilst this response cannot pre-empt the results of the Strategic Defence Review (SDR), which will have a significant impact on how the Department takes forward work against several of the Committee’s recommendations, the Department continues to make progress enhancing AI adoption and seizing delivery opportunities in parallel.

7. The MOD is fully committed to accelerating our safe and ambitious adoption of AI, and this is manifesting in several ways. Our commitment to promoting and fostering AI skills and talent extends beyond internal development, as the Department seeks to remove barriers for our industry partners, facilitating collaboration and innovation. By fostering a strong partnership between the MOD and industry, we aim to harness the collective expertise and resources to drive progress in the field of AI. The MOD is dedicated to advancing both technical and policy enablers to create an environment that supports the responsible and effective use of AI technologies. The MOD also plans to develop a more integrated and dynamic relationship with investors, industry, and research partners to provide a strong, sovereign ecosystem for future AI-exploitation, including through DIS.

8. Furthermore, the MOD recognises that the impacts of AI transcend borders. The MOD plays a proactive role in international regulatory diplomacy to ensure ethical considerations and the associated apparatus of policies, process and doctrine facilitate the legitimate, responsible, and ethical development of AI; as well as driving forward our efforts to enhance collaboration and interoperability with key partners. This will ensure that we put robust and battle-winning AI-enabled systems into the hands of our users to outpace our adversaries. The UK is a leading voice in international dialogues around the use of AI and autonomy in Defence. Through existing guiderails and frameworks already well established in the defence sector and through International Humanitarian Law (IHL), we will develop and deploy AI-enabled systems for purposes that are demonstrably beneficial: driving operational improvements, supporting the Defence Purpose, and upholding human rights and democratic values. We will work with allies to accelerate collective understanding of these technologies through mechanisms such as AUKUS. Through these efforts we will position the MOD at the forefront of AI innovation, enabling us to meet the evolving challenges of the modern Defence landscape while upholding the highest standards of ethics, safety, and accountability.

Section 1 – The UK’s Defence AI Landscape

9. This section addresses one general conclusion and recommendation from the Committee’s report on the UK’s Defence AI Landscape, including points about international comparisons.

10. The Government agrees with the Committee that it is important to compare the UK’s Defence AI sector against international peers to learn lessons and target interventions. The MOD has undertaken early research to assess the scale and health of the Military AI industrial ecosystem, including limited benchmarking against other nations. This has informed international engagement, strategy development and briefings and the MOD is exploring options to expand on this initial analysis, although it is unlikely that any final products could be made publicly available.

11. The MOD also works closely with cross-government partners to understand, support and where appropriate catalyse the wider UK AI technology sector. As highlighted in the recent HMG AI Opportunities Action Plan, published by DSIT, the UK is home to a thriving AI sector and recognised as a world-class hub for tech start-ups, research, innovation, and digital industries. Areas of particular strength include fintech, biotech, digital marketing & advertising, and security. The UK’s position as a global AI leader is further strengthened by its respected legal system, well-developed regulatory approaches, research powerhouses like the Alan Turing Institute and influential Government bodies such as the AI Safety Institute. The MOD’s Defence AI Centre (DAIC) actively works with partners to leverage skills and capabilities across this ecosystem, to break down barriers to engagement and encourage deeper and stronger technical collaborations.

Section 2 – The Ministry of Defence and AI

12. This section addresses several conclusions and recommendations from the Committee’s report, specifically on the section covering the Ministry of Defence and AI, including points about the Strategic Defence Review (SDR), technological developments in Defence AI, and becoming ‘AI ready’.

13. The Government is committed to taking bold action to fully realise the transformative potential of AI in Defence, and agrees in general terms with the recommendation – although this response cannot pre-empt the findings of the SDR. The Department has established a track record of delivering quality and productivity improvements, new capabilities, and decision support tools using AI and automation. Now the Department has a unique opportunity. The SDR and related Departmental processes such as Defence Reform will consider structures, authorities, accountabilities, resources, and delivery mechanisms, all of which are very relevant to achieving AI aspirations. They provide an opportunity to conduct rigorous strategic planning and prioritisation, ensuring Defence operates as a cohesive and integrated end-to-end system, aligning Defence’s capabilities to respond to the current and future threat landscape. Whilst the results of the SDR are being finalised, we will ensure the Department considers the Committee’s recommendation when planning SDR implementation. The Secretary of State’s publicly announced Defence Reform changes to the organisational structure for the MOD—whereby the organisation will coalesce under a new top team of PUS, CDS, NAD, CDN—is a first step on creating an organisation that can rapidly identify and adopt new technologies.

14. The Government recognises that there have been rapid technological advancements and evolving circumstances since the Strategy’s publication, and we have taken significant steps to adapt and respond to these changes. As we await the outcome of SDR, we anticipate the need to refresh our strategic approach to AI to align with the Departments broader strategic vision. Consequently, while we appreciate the Committee’s interest in the details of 2022 Strategy implementation, we believe it is more pertinent to focus on shaping our future approach. However, we agree with the Committee that specific, measurable actions will be essential to tracking progress and intend to ensure this is integral to our approach.

15. Since 2022, the most notable technological development has been the advent of General-Purpose Transformer (GPT) architectures. This breakthrough has galvanized an explosion of interest in AI, largely due to a) the order of magnitude improvements in the outcomes generated by such models and b) the method of consumption, as the models are incredibly easy to access through simple chatbot interfaces.

16. In the operational space, innovations in Ukraine have been notable. UAVs have shifted from a niche capability only available to a handful of nations, to a commoditised capability able to perform autonomous operations in a denied environment. The UK has led research into the development and understanding of vulnerabilities of such platforms, remaining clear we oppose the creation and use of systems that would operate without meaningful human control, achieved through context-appropriate human involvement throughout their lifecycle. Progress has been made in digitising and bringing an element of automation and controlled autonomy to the intelligence enterprise and to Command and Control across different domains.

17. Aside from the obvious benefits in terms of capability improvement, these breakthroughs have had multiple effects on the implementation of the AI Strategy. Strategically, the widespread utility of this technology has underscored that AI is not a fringe technology that promises incremental gains in efficiency and effectiveness, but a truly revolutionary enabler that must be harnessed across the enterprise if the MOD is to remain relevant. This has resulted in growing support amongst Defence senior leaders, embodied by the establishment of the 2PUS-chaired AI Delivery Group which seeks to drive coherent progress across the enterprise. The experience of the past two years has also underscored the need for a nimble approach to research, development and delivery where the proliferation of AI-enabled capabilities available through vendors and open source has necessitated rapid efforts to update Defence policy and ways of working to mitigate unforeseen vulnerabilities while realising the promise of these applications.

18. Looking forward, we are seeking to further build on working relationships across DSIT and the Defence and National Security ecosystem to realise economies of scale and opportunities to share resources that support our engagement with industry, growth of skills and talent and potentially model development.

19. The Government agrees that becoming ‘AI Ready’ should be an integral part of its broader transformation as we seek to drive cultural change in Defence. We are also interested in the suggestion in the report that the Department should move towards the language of becoming ‘AI Native’ to reflect our intention to embed the skills, enablers, and cultural change required to cement AI’s position as a key component of the Defence toolkit. The scale of change Defence faces is unprecedented, requiring a fundamental shift in how we approach technology, innovation, skills, culture, and our relationship with industry. This is not just a technological challenge, but one that requires a systems approach to ensure our people are prepared to adopt new AI technologies to secure Defence advantage. We are fully committed to using the outcome of the SDR to accelerate our efforts to embed AI in Defence.

20. Transforming the MOD into an AI Ready organisation is a seismic change and one that will require joint efforts from all Top-Level Budgets and functions, and as such the Department has viewed AI Readiness as a central part of its efforts to develop AI in every part of our business. Alongside establishing a range of initiatives which will benefit our AI programme, such as the Digital Skills for Defence (DS4D) programme, which are central to our digital transformation efforts, Defence has taken a number of specific actions to drive cultural change and AI Readiness, including:

• The 2nd Permanent Secretary wrote to all Defence leaders last year to set out the need for senior leads to go further to accelerate AI adoption and shift our way of thinking towards an ‘AI by Default’ approach. Senior leaders in Defence have also taken part in AI focused events via the Digital Leaders Learning Programme (DLLP), which identified key opportunities for the transformation of Defence business using AI and highlighted a set of collective challenges to be overcome to fully realise those opportunities.
• As part of implementing the Dependable AI JSP 936, Defence organisations have nominated a Responsible AI Senior Officer. A part of this role is to be a senior champion who drives the safe experimentation with AI in their organisation, removing barriers and overseeing top-level risks and benefits, and driving a pro-ethical institutional culture. The argument for creating a culture that fosters responsible AI adoption and methods for enabling this is compelling. The DAIC, the MOD’s AI Ethics Advisory Panel and the AI Expert Group have been leveraged to support training for these senior leaders.

21. The Government agrees CDS has a pivotal role in maximising the UK’s strategic military advantage and leading Armed Forces transformation. Digital, data and AI will underpin every aspect of the MOD’s operating model moving forward. The SDR and Defence reform offer timely opportunities to put in place the structures and levers required to enable this transformation and drive greater cross-domain capability integration. Under the new model announced by the Secretary of State, CDS will formally command the individual Service Chiefs for the first time, ensuring faster delivery and clearer accountability across Defence while driving transformation, coherence and interoperability. The Military Strategic HQ (MSHQ) will have a critical role to play the development of concepts, doctrines and plans against a pacing threat, ensuring digital, data and AI are central to the way our Armed Forces operate and fight. The new model will have the Military Strategic Headquarters articulate broad “Problem Statements” (i.e. military effects that they seek to achieve) that the new National Armaments Director will work with industry in a rapid, integrated fashion to propose solutions to. This will ensure the MOD leverages the cutting-edge innovations that industry can bring, and not be constrained by historical issues of over-specification and constant changes to specifications. The Government agrees with the committee that this is about driving cultural change to make operating in this way “business as usual”, and the MSHQ will drive this supported by Strategic Command. However, evolving Defence to be an AI-native organisation is a collective endeavour; senior leaders across the NAD and DoS will play a critical role in accelerating the widespread exploitation of AI.

22. The Government agrees that the DAIC must be accessible to the spectrum of industry and the UK AI talent base. This does not necessarily need to be achieved through a single physical focal point. The DAIC currently has a considerable presence across three core hubs and three satellite locations enabling access to a wider swathe of talent across the UK while providing the flexible working arrangements expected in the modern tech-sector. A strong presence across these locations has fostered effective collaboration leading to projects migrating from the science and technology enterprise, through innovation teams into the hands of users at scale via the DAIC.

23. On talent exchanges and placements, the Government fully agrees – the DAIC is currently hosting its first secondee from industry and has established military posts for five additional military colleagues with a blend of technical and operational experience. It is exploring the establishment of bi-directional secondment programmes with adjacent departments as well as private sector partners. Building on these early efforts, DAIC intends to continue to develop its workforce strategy, including consideration of the use of Reserves structures, in order to make increasing use of the talent industry has to offer.

Section 3 – Developing Defence AI

24. This section addresses several conclusions and recommendations from the Committee’s report, specifically on the section covering Developing Defence AI, including points about the Integrated Procurement Model, working with SMEs, and developing contracting processes.

25. The Government agrees on the importance of expert advice, including from industry, within procurement. A key element of the Department’s acquisition reforms is early expert advice and engagement with industry on factors such as technology, to inform decisions on possible solutions and enable us to take advantage of the innovations the market has to offer. Programmes are encouraged to routinely consider the potential benefits of AI-enabled solutions when developing their requirements. The Department is developing mechanisms to support programmes in considering and balancing broader policy priorities, including around the use of technology such as AI.

26. The Defence Industrial Strategy Statement of Intent published in December 2024 sets out the intention to “foster a more diverse community of suppliers, including nontraditional Small and Medium Sized Enterprises (SMEs) and future technology”.

27. In addition, the MOD plans to shortly launch the ‘Defence Tech Scaler’ (DTS) which is an initiative being run by Defence Digital commercial. The DTS will establish an end-to-end process for how the MOD engages, nurtures and grows software, Data and AI companies from new suppliers, through to firmly established major suppliers. The purpose is to set up suppliers to be Digital Backbone compliant from the outset, enabling potential future growth and scalability. This will support supplier on-boarding, provide access to sponsorship for clearances and will also allow vendor access to MOD data. DTS is not just a commercial process, it encompasses improved and coherent involvement from Technology, Finance, Cyber, Commercial and Strategy. This will support the prosperity of Defence suppliers and, critically, ensure that suppliers are Digital Backbone compliant.

• Sending a clear demand signal to those companies about what kind of AI it is interested in acquiring, and clearly signpost how smaller companies—including those who have never worked with the MOD before—can access opportunities to sell to the department;
• sending a clearer signal to the capital markets that it takes developing an ecosystem of UK AI companies seriously, to increase investor appetite for the sector;
• sponsoring promising early stage research and development with potential defence applications, to help develop proofs of concept while building UK skills and industry; and,
• developing ways in which start-ups and SMEs can partner with defence primes to build their capacity and expertise. (Paragraph 66)

28. The Government agrees that it should cultivate a more dynamic ecosystem of smaller AI companies working in Defence to accelerate MOD AI Procurement and deployment through private sector agility, often at the forefront of AI. The MOD also agrees with the specific recommendations for cultivating relationships with smaller AI companies. The new National Armaments Director will be accountable for shaping an indigenous market for AI companies that can support the MOD on delivering its ambitions.

29. Sending a clear demand signal to those companies about what kind of AI it is interested in acquiring: It is necessary to set out MOD expectations for AI-enabled capabilities. By aligning with Defence needs, all companies, including SMEs and non-traditional Defence suppliers can contribute to building the required AI capabilities:

• The Defence AI Playbook, developed by DAIC, outlines enduring AI challenges and opportunities within the MOD. An updated version will incorporate industry feedback and provide guidance on effective collaboration and navigating commercial frameworks.
• DAIC Connect is an industry engagement day aimed at collaboration within the AI community, signalling Defence demands. This includes workshops between AI industry representatives and defence end users on how and where we can collaborate innovatively to accelerate adoption. This facilitates deeper engagement beyond short-term contracts, promoting sector wide growth and understanding.
• The Dependable AI (JSP 936, published on gov.uk) sets out both a demand signal and clarity on our Responsible AI expectations for partners, detailing how the MOD’s AI Ethical principles must underpin the delivery of robust, reliable and effective AI-enabled services and capabilities. The DAIC also leverages an AI Expert Group to help refine policy directives in line with technological advancement and evolving AI adoption.

30. Clearly signpost how smaller companies can access opportunities to sell to the Department: It is critical to develop ways in which smaller, non-traditional AI companies and SMEs can access ways to showcase their technology and sell to the Department. There have already been key successes in this area including:

• Defence Digital Commercial owns AI as a ‘Commercial Category’, in line with the business function strategy of Category Management, partnering with ‘risk embracing’ Commercial Cells like CommercialX. As the main delivery arm for DAIC, initiatives such Dragons’ Den have boosted opportunities for smaller companies to sell and showcase their technology. The reoccurring ‘Tech Showcases’ workstream, targeted for launch later in 2025, will deliver open pitching sessions for small companies to present to defence personnel.
• The Defence and Security Accelerator (DASA) has partnered with IoT Tribe to provide business acceleration services from its London hub. This collaboration, named ‘Janus’, will help smaller companies develop dual-use technologies for both defence and commercial markets, enhancing the UK’s role in NATO’s Defence Innovation Accelerator for the North Atlantic (DIANA) initiative. DIANA provides funding and expertise to develop start-ups working on emerging and innovative technologies that have a dual use, in both the civilian and defence markets. This offers further opportunities for UK companies to grow and access markets across the NATO Alliance, and to develop their products to meet defence needs.

31. Send a clear signal to capital markets that it takes developing an ecosystem of UK AI companies seriously, to increase investor appetite: Given most AI developments are underpinned by market dynamics and private sector investment, we agree with the need to engage capital markets, particularly the Venture Capital (VC) space. This was articulated clearly by the Secretary of State at the London Defence Conference, hosted by Rothschild, where he announced the Defence Industrial Strategy.

32. Initial discussions between DAIC colleagues and representatives from capital markets have taken place, and we are in the early stages of planning a dedicated stream of activity focused on the investor community. This initiative aims to provide investors with a clear understanding of Defence challenges in the medium-term, empowering them to make well-informed decisions regarding capital allocation.

33. Sponsoring early-stage research and development with potential defence applications, to help develop proofs of concept while building UK skills and industry: Defence recognises the importance of this and has made good progress in this area, with a specific example in DAIC-X’s DASA AI innovation:

• DAIC-X supported proposals for a DASA AI innovation focus area, which cultivated innovative research projects that applied AI to Defence challenges or aimed to overcome common barriers to implementing AI within Defence. 10 projects were funded (up to £150,000 each) across 10 companies, including 3 small and 2 micro-enterprises. Half of these projects have been carried forward by the DAIC, the RAF or the wider S&T programme.

34. Developing ways in which start-ups and SMEs can partner with Defence primes to build their capacity and expertise: Moving forward, it will be important to ensure that SMEs and Micros are involved throughout the entirety of a program’s life cycle and primes used to accelerate scalability across the organisation. Defence values collaboration through Commercial initiatives amongst Primes, SMEs, Micros and academia. For example, pairing suppliers with similar capabilities to facilitate SMEs that may not be able to bid independently to have an opportunity partner with other organisations which can enable their scalability.

35. The Committee will be aware of the announcements made on 28 February and 3 March of steps Defence will take to turbocharge innovation and boost small businesses benefitting from UK’s defence investment. We will establish a new Defence innovation organisation under the new National Armaments Director; launch a new hub to provide small and medium enterprises (SMEs) with better access to the Defence supply chain; and set direct SME spending targets for the Ministry of Defence by June this year. These moves are designed to boost the resilience of the UK’s defence industrial base and to seize the full power of rapid technological change as part of the Government’s forthcoming Defence Industrial Strategy.

36. The Government agrees with the Committee’s recommendation and recognises the need for a cultural shift in defence acquisition. We have introduced tools to help leaders and teams to understand the risk and complexity of their programmes, empowering them to use their professional judgement to take risk-based decisions to drive pace. A Psychological Safety Assessment Service is available to our major programmes to help leaders create the right environment for successful delivery in their teams, where people feel empowered to raise ideas and to innovate and are confident exposing risks and issues early so that these can be addressed.

37. The MOD can drive this cultural shift by embedding flexibility and innovation within the procurement cycle. CommercialX have seen success through streamlining commercial pathways and including break clauses at key decision gateways, enabling faster procurement processes, and ultimately allowing the decision makers to pivot quickly if needed. ‘Market Engagement by Default’ is the current operating model for the AI CommercialX team. Through better intelligence, it has ensured decision makers have a clearer understanding of technological possibilities, risks, and trade-offs – this is evidence in the case studies highlighted under Recommendation 12. In addition, CommercialX has developed a management development programme that is accessible to all commercial officers and touches on soft skills needed to address the culture gap. It is looking to build innovation mindsets, increase resilience, and promote a risk aware and business savvy commercial function. It has been designed specifically to modernise and build new ways of working across all the levels of our workforce.

38. The Government agrees with the Committee on the importance of Spiral acquisition and is promoting greater use of Spiral approaches, supported by updated guidance and a new Spiral Commercial Pathway. The AI Category Team and CommercialX are key contributors to this activity, sharing their case studies. Increased use of Spiral acquisition approaches enables innovative solutions, with faster delivery of a minimum deployable capability that can then be iterated to reflect learning from user experience and advances in technology. A circa 200-strong Spiral Community of Interest has been established and is sharing experiences and learning to shift culture and behavioural norms toward new ways of working.

39. The Government agrees that it should strive to pull through successful defence innovations (including those which use AI) into fielded and scaled capabilities. Defence’s acquisition reform should facilitate more regular development, deployment, and integration of innovations into defence systems. Updated departmental guidance and pathways will enhance support for spiral development and digital technologies department-wide. With its federated structure, the DAIC plays a crucial role in accelerating and scaling AI projects emerging from the innovation enterprise as well as providing common AI Services that enable exploitation at scale.

40. In supporting and enabling procurement of innovation across Defence, the Department has various programmes of funding which are targeted to deliver, exploit or scale defence innovation including:

• Defence Innovation Fund (DIF): which aims to enable, generate, and put innovation ideas into practice for Defence.
• Research and Development (R&D) Exploitation Engine (ExE): which aims to pull-through promising R&D into capability for Defence.
• Centralised R&D funding: which aims to support R&D, including Spearheads and Game Changers, ensuring a continuous flow of technology development to be exploited for Defence in priority capability areas.
• Test and Evaluation (T&E) Transformation Fund: which aims to support innovation across the Department to explore and adopt new ways of thinking about, conducting and utilising T&E, including for innovative or novel technologies.

41. Ongoing discussions between the Department and HM Treasury regarding the Department’s future budgets prevent a direct response to the Committee’s recommendation to establish a new fund. However, the Department acknowledges the Committee’s recommendation on this matter and will incorporate it into its ongoing discussions with HM Treasury where it is appropriate to do so.

42. The Government agrees with this recommendation, acknowledging the potential administrative barriers in entry to contracting processes for AI. Defence AI procurements will be undertaken under the stewardship of Defence Digital Commercial (who partner with CommercialX). By adopting category management Defence Digital and CommercialX will own and direct commercial strategies that are fit for purpose for digital labilities. It will reduce administrative barriers, help internal and external coordination of capabilities, drive education and good practice and increase market opportunities by connecting the dots across the enterprise. The AI Category in Defence is taking the lead across government by developing routes to market, sharing guidance, best practice and commercial vehicles that are designed to be reusable, all whilst providing a consistent approach for SMEs and realisation of the benefits of a simple approach.

43. CommercialX has developed and is using simplified documents across a range of subjects that help SMEs in their engagement with Defence. These include a set of terms and conditions that have been cocreated with SMEs, written in accessible language (rather than using legal or Defence-specific terminology) for use in low-value, simple procurements; guidance on routes into the MOD, for internal staff, and ‘myth-busting’ activities around early market engagement. Again, while not specific to the AI sector, these support CommercialX’s ambition to reduce barriers to entry for SMEs.

44. CommercialX has developed language that can be added to tender notices that will allow solutions to scale by default without the need to re-run competitions. This language that can be added to any digital tender that will allow low-value procurement to be run once and allow that capability to scale across the enterprise when further demand is identified, up to £50 million.

45. ‘How to Buy AI’ Guidance has been created and distributed internally, targeted at those involved in purchasing AI-enabled solutions. Through this guidance the category team is focusing on disseminating knowledge and lessons learned via iterative updates and best practice playbooks which will facilitate the consistency and standardisation of this approach. This knowledge is also shared more broadly through the MOD’s contributions to the Government Commercial Functions AI Working Group to broaden the benefits.

46. The category team also look at novel contracting solutions and share best practice through the above. For example, CommercialX has had great success using the Home Office’s Accelerated Capability Environment (ACE) to enable alternative suppliers to easily access AI contracting opportunities within the Royal Navy. Case Study 1: Proving the value of the Royal Navy’s AI roadmap – Case study – GOV.UK Case Study 2: Using data to predict and support the future Royal Navy – Case study – GOV.UK. Both show examples of UK SMEs working alongside either academia or larger defence suppliers to deliver defence outputs.

Section 4 – Data and infrastructure

47. This section addresses several conclusions and recommendations from the Committee’s report, specifically on the section covering data infrastructure, including points about the resilience of our digital ecosystem, and the Digital Strategy for Defence.

48. The Government agrees with the points put forward by the Committee and is in the process of refreshing the mapping of our digital ecosystem for the Digital Backbone against our Technical Reference Model, which was jointly developed with Industry and partners. We are using this to identify gaps, duplication and areas that need further investment or development.

49. The Government agrees with this recommendation and the Digital Backbone remains a core concept for UK Defence.

50. The Digital Backbone is an interconnected digital ecosystem that integrates data, applications, and systems across multiple domains, partners, allies, and suppliers. This ecosystem supports both departmental business and war fighters by leveraging cloud services, the Internet of Things, and next-generation networks. It provides accessible services and data to users when and where they need it, ensuring seamless and secure information sharing for strategic decision-making.

51. The Digital Backbone exists today, facilitating rapid development and deployment of new applications and services across its common services layer, continuously improving to enhance quality, throughput, reuse, functionality, reach, data access, efficiency, and cost management. The Digital Backbone will continue to evolve, upgrade, mature and flex to respond to changing threats and needs of the users across Defence. It is a ‘living’ thing that has no ‘completion’ nor ‘end’ date. Delivery of core technology elements of the Digital Backbone are closely measured by Defence Digital, which includes the Digital Programmes in the Government Major Projects Portfolio.

52. Recognising that scalability, interoperability and long-term value are essential for the enablement of the Digital Backbone, the AI Commercial team have focused their strategy around ‘do-once, use-many’ scalable contractual agreements, both through the technical readiness levels and access across the organisation. The ‘Defence Tech Scaler’ (as mentioned in recommendation 8) is the next step for this strategic workstream. This mechanism will enable delivery teams to access Digital Backbone compliant solutions through standardised interoperable products, Data and Cybersecurity alignment and commercial flexibility whilst enabling MOD to maintain Strategic Supplier Relationships with key technology partners, which will ensure sustained investment in the digital infrastructure and continued innovation and ensure Value for Money.

53. The Government partially agrees to this recommendation and is committed to the strategic direction set out in the Digital Strategy. Through the Strategic Defence Review (SDR), Spending Review, and Defence Reform (DR) transformation, government is making significant changes to transform the way Defence decides, directs, and delivers. The implementation of the Digital Strategy, as well as other strategies, will need to reflect the outcomes and priorities in SDR, Spending Review and DR.

54. Consequently, Defence is planning to develop a refreshed digital strategy fully aligned underneath Defence Review and Defence Reform implementation headmarks and reflecting the revised Defence Strategy and Plan. It will be underpinned by an implementation plan with associated metrics to accurately measure progress. The Department continues to drive delivery of the Strategic Outcomes (SO) in the Digital Strategy, including various measures of success. A high-level update on progress is set out below.

55. Update on Strategic Outcomes (SO) in the Digital Strategy:

• SO1 – Exploit Game Changers and Data at pace and scale via the Foundry. Foundry and DAIC are operational and leveraging key datasets to enable data driven insights and efficiencies throughout the department, e.g. deploying AI to support the SDR process. Demand for services exceeds current supply, so a prioritisation exercise in ongoing, informing SDR proposals which seek increased resource to meet demand.
• SO2 – Deliver and adopt technology to support the concept of the Digital Backbone. The Digital Identity service has been launched, with a roll out underway with Multifactor Authentication (MFA) being prioritised. On-premise Secret hosting is now available through Virtual Infrastructure Exchange (VIX) with 188 live services migrated. MOD Community Cloud@S (hyperscale cloud) has been delayed due to vendor withdrawal, and the Invitation to Negotiate (ITN) expected to be released in 2025. At the 2023 Coalition Warrior Interoperability Exercise (CWIX) the NATO core data framework capability was exercised and demonstrated the ability to share information using common standards driving interoperability.
• SO3 – Reset Cyber Defence to materially lower risk. The Secure by Design concept has been embedded across Defence, ensuring security is an integrated part of capabilities from the beginning of any project. The cyber compliance scores continue to improve and there is additional work underway to address obsolescence and strengthen the wider digital and cyber ecosystem.
• SO4 – Enable a step change in digital delivery. The milestones for achieving a material modernisation of corporate services to Defence that are relevant, reliable, and embracing digital are on track to be delivered. The service executive model has been established and the process embedded. The CommercialX pilot has executed 147 of 205 contracts ahead of target award date and delivered £47.14 million of savings against total contract values of £328.6 million and is achieving a c30% increase in efficiency compared to traditional Defence acquisition benchmarks. In 2024, the pilot began transitioning to an enduring capability.
• SO5 – Routinely develop and access leading-edge digital talent. Successes against this include the induction of 100 Students into the STEM and Digital Bursary Scheme, multiple Senior Leaders Digital Masterclasses, digital skills accelerator courses and a new Digital Skills Learning Platform to increase access for the whole 250,000 workforce (civilian, reserves and regulars) to digital learning. There have been recruitment controls in place, but our delta for skills required is driving a fundamental rethink on how we recruit.
• SO6 – Embed a unified function enabling pan-defence cohesion. Increased visibility across the digital portfolio has resulted in enhanced strategic alignment of Top-Level Budget (TLB) plans, resulting in efficiency targets being on track to be delivered by 2025. Proposals to enhance functional levers and authorities to cohere the pan defence Digital Portfolio have been submitted as part of the SDR process.

56. The Government agrees with the recommendation to address barriers to AI SMEs and data accessibility. Smaller and micro-organisations are often thought leaders in this sector. To establish mutually successful relationships that protect and benefit all parties, new thinking and innovative commercial solutions are required. The DAIC is leveraging forums such as the AI Expert Group and the new Defence Industrial Joint Council to take views directly from industry members (including SMEs) on barriers to engaging with the MOD, including on the challenges around accessing relevant data sets. The MOD will work to address these barriers though the following initiatives and next steps:

• CommercialX and the AI Commercial Team within the MOD have recently created internal guidance and an e-learning module which aims to support commercial and non-commercial staff in identifying and resolving critical considerations when an AI-enabled solution is proposed. A key area of focus is data; how it is sourced and validated, how it is made accessible to suppliers, and how AI-informed outputs are used. A subsequent iteration that has commercial-specific content at its core, e.g. commercial, exploitation and intellectual property levers, is planned if funding permits.
• In conjunction with the AI Category Team, CommercialX will develop commercial approaches and solutions that form the commercial blueprint for wider Defence to optimise access, and enable scaling, using open-source, synthetic and Defence-specific datasets as they are available. These include the creation of data sets for use under the ACE framework and developing a UK version of CRADA (Co-operative Research and Development Agreement).
• CommercialX has run two “Dragons Den” events where selected SMEs were invited to present their responses to a challenge set by a delivery team to a panel and audience comprising wider Defence stakeholders and internal and external investors. Six of the total fourteen suppliers that took part in the event have gone on to contract with defence for numerous capabilities.
• CommercialX has developed and is using simplified documents across a range of subjects that help SMEs in their engagement with Defence. These include a set of terms and conditions that have been cocreated with SME’s, written in accessible language (rather than using legal or Defence-specific terminology) for use in low-value, simple procurements; guidance on routes into the MOD, and for internal staff, ‘myth-busting’ activities around early market engagement. Again, while not specific to the AI sector, these support CommercialX’s ambition to reduce barriers to entry for SMEs.

57. For SMEs to access and use high-security Defence data, Defence must comply with policy for National Security Vetting owned by the Cabinet Office, and Defence security clearances for individuals using data would be delivered by UK Security Vetting (UKSV), within the Cabinet Office. Performance has continued to improve and at the end of December 2024, 97.5% of SC application for Defence cases, including contractors, were completed within 25 days, which is a 90% improvement from February 2023.

Section 5 – Skills and the workforce

58. This section addresses several conclusions and recommendations from the Committee’s report, specifically on skills and talent.

59. The Government agrees that Defence should understand the skills it needs and how they will continue to evolve. The MOD recognises that equipping the Armed Forces and the wider Defence workforce with the skills to use, understand, and integrate AI technologies is essential for harnessing AI’s transformative potential safely and responsibly. As such, the Department has begun conducting the AI skills mapping exercise. This is taking place in parallel to the wider process of identifying the skills that are present within through the Pan Defence Skills Framework (PDSF) which is developing a common skills taxonomy that will initially allow the MOD to assess our organisation using a single benchmark or metric.

60. The DAIC has conducted an initial, high-level analysis of the likely main categories of personnel across the defence workforce and the AI skills they will likely need in the near-term. This analysis centres around five AI Personas: AI Explorer, AI Warfighter, AI Business Operator, AI Professional and AI Senior Leader. These personas recognise that the skills needed in Defence will be diverse, encompassing a small cadre of AI experts able to design, build and extend AI systems, a larger cohort of practitioners able to fully exploit AI systems for defence purposes, and finally a common baseline skill level for all staff which ensures all can use AI tools safely and responsibly.

61. Furthermore, the Dstl Future Workforce and Training programme will undertake a research project on the People Implications of AI, running over the next 5 years – this might be leveraged to inform a future skills mapping exercise. The Dstl research project aims to understand the people implications to ensure that the MOD acts as an intelligent customer for AI adoption with respect to personnel, training, and human factors. To help the MOD address current and future challenges related to these topics, the project is anticipated to focus on providing evidenced advice on the effect of AI on workforce size and composition, harnessing skills for accelerated AI adoption, and optimising the human factors associated with AI adoption.

62. The Government agrees, and the Committee should note that MOD is currently developing a new plan for AI Skills and Talent. This plan identifies “AI Professionals” as a key persona (see above) where we need to focus our upskilling and talent efforts. These are the experts we need to research, build, develop and maintain the defence AI systems of the present and the future. The MOD is committed to creating opportunities and careers for AI Professionals within the MOD and Armed Forces. The scope of this work includes: establishing schemes that enable AI Professionals to move between the defence and civilian sectors flexibly, including industry placements; adopting a ‘Build, Borrow, Buy’ methodology to harness talent from inside and outside defence; and, incubating and developing talent through multiple methods for both military and civil service personnel.

63. While our thinking is at an early stage, Defence has already begun work on the first phases of our plan:

• Creating an AI talent roadmap: Understanding the likely numbers and skills required and identifying development schemes and on-ramps to upskill, attract and retain AI talent, likely to include apprenticeships, graduate and post-graduate programs, external hires, secondees, and reserves. An example would be Defence Digital hosting a large cohort of graduates, including placements in the Defence AI Centre.
• Exploring collaboration—for example shared training and staff interchange—with other Government organisations with similar requirements for AI Professionals.

64. More broadly, Defence is pivoting to a skills-based enterprise that values people as individuals, celebrates their collective talent, knowledge, and skills, and invests in them as a national capability. The MOD recognises that growing more digital and cyber skills, establishing digitally connected talent pools with a common skills framework that covers industry, academia, other government agencies and our alumni of Defence veterans is important for enabling zig-zag careers between Defence and civilian sectors. In respect of specialist recruitment, Defence will consider multiple avenues for meeting the requirement for AI talent, including harnessing the technical expertise of our reservists. The MOD is already taking steps to attract and retain talent in a challenging marketplace for AI skills, for example by implementing the Digital Skills Allowance on top of base civil service salaries. This is now a reality, with the first critical AI-related roles in the MOD attracting this pay uplift as of December 2024.

65. The Government agrees with this recommendation and is committed to equipping defence operators and leaders with the knowledge and skills to understand and leverage AI technologies effectively. This will ensure that Defence is well-prepared to harness AI’s impact and maintain a competitive edge in defence operations and strategic decision-making.

66. As mentioned in response to recommendation 17 (Paragraph 61), the skills effort is organised around five AI learning personas, based on industry analogues but tailored to defence. Initial work to support our AI Senior Leaders and AI Warfighters has included:

• Holding a supplier-led AI Immersion Day in MOD HQ aimed at AI Senior Leaders.
• Working with the Joint Services Command and Staff College at the Defence Academy to increase the proportion of professional military education dedicated to AI. One key initiative is the Train-The-Trainer AI pilot, supporting the integration of AI training into the syllabus through four pilot studies. The intent is to equip learning and development staff to design and implement Operational Planning exercises involving the use of AI tools, as part of the flagship Command and Staff courses for all military Officers.
• Developing our initial contextualised (defence-specific) AI learning for war fighters and senior leaders.

Section 6 – Cooperation with Allies and AUKUS Pillar 2

67. This section addresses several conclusions and recommendations from the Committee’s report, specifically on the section covering cooperation with Allies and AUKUS, including points on the alignment with AUKUS and NATO allies.

68. The Government recognises that an understanding of Defence AI objectives and strategies amongst our allies within AUKUS and NATO is of key importance. The precise nature of discussion will need to be determined following the SDR and will need to reflect new responsibilities and accountabilities established under Defence Reform. AUKUS is the most significant defence and security partnership the UK has entered in the past 60 years and the UK Government is committed to maximising the benefits of the partnership. Through AUKUS Pillar 2, our trilateral science and technology, acquisition and sustainment, and operational communities are working across the full spectrum of capability development – generating requirements, co-developing new systems, deepening industrial base collaboration, and bolstering our innovation ecosystems.

69. Through AUKUS Pillar 2, the Government is collaborating with the US and Australia to accelerate collective understanding of AI and autonomy technologies, and how to rapidly field robust, trustworthy AI and autonomy in complex operations while adhering to international law and ensuring the ensuring safe and responsible use of AI. This work includes the development of common standards and architectures for AI and Data, and common frameworks for Test and Evaluation of AI. These will help to reduce barriers between our national AI ecosystems. In October 2023, the AI TORVICE trial was held which sought to accelerate the adoption and understanding of AI and autonomous systems. This was preceded by the first AUKUS AI trial held in April 2023, which achieved world firsts, including the live retraining of models in flight and the interchange of AI models between AUKUS nations.

70. The Government is also a leading contributor within the NATO Data and Artificial Intelligence Review Board (DARB). The DARB is tasked with governing the responsible development and use of AI in accordance with NATO’s Principles of Responsible Use. Its primary objectives include building trust, guiding the adoption of responsible AI, and acting as a forum for the exchange of information among Allies. As a trailblazer in Responsible Military AI use and adoption, the UK has actively contributed to the development and piloting of the NATO Responsible AI (RAI) toolkit, participated in tabletop exercises, helped to shape Alliance thinking about ‘AI readiness’, and supported the development of a test bench for validation, verification, and vulnerability assessments.

71. We have been central to the formulation of a common, NATO-wide strategic approach to new technologies, including AI and the common enablers required for its rollout. The UK was one of the leading voices calling for NATO to revise its AI Strategy ahead of schedule last year – reflecting the pace of technological change. We are working with the Alliance to improve NATO’s facilitation of technology adoption through procurement initiatives, operational experimentation, and doctrinal and cultural change.

72. The Government is also playing an active role in other multilateral fora which include AUKUS and NATO partners. Through the Five Eyes (FVEY) Combined Digital Leaders Summit, the UK is aligning its approaches with partners on AI assurance and governance through workshops and pursuing data interoperability and integration through working groups. As a core member of the 16 nation AI Partnership for Defence community, the UK is collaborating with AUKUS and key NATO partners to share lessons on the building blocks for responsible, ambitious AI adoption. The UK shows leadership in this forum, regularly promoting our approach and interoperability with allies.

73. Work is underway to establish how AI-enabled applications developed by AUKUS partners can be more readily shared, enabling accelerated transfer of knowledge, skills and industry developed services that support the UK and allies in maintaining their strategic advantage, building skills and capacity across our organisations.

74. The Government agrees on the importance of collaboration with the US and Australian Governments in integrating research and development in AI between the UK and AUKUS allies. Under AUKUS Pillar 2, the UK signed a Project Agreement (PA) with the US and Australia in November 2023 to collaborate on AI and Autonomy Research and Development. The AUKUS Resilient and Autonomous Artificial Intelligence Technologies (RAAIT) PA is a lab-to-lab agreement which has enabled joint and collaborative research, development, and experimentation, including trials noted in the report (para 110). This agreement continues to strengthen AI R&D collaboration between the three nations and their industry partners.