Introduction: Artificial Intelligence (AI) poses an intrinsic threat to the UK manufacturing sector.
The UK and particularly, the manufacturing sector, needs to understand AI's impact on productivity, employment, skills, and international competitiveness. This needs to happen in an empirical framework to govern the actionable, descriptive measure of employability demanded by AI adoption in manufacturing. The most relevant, applicable literature on the theme is thus surveyed to construct an empirical and policy-oriented assessment.
The key questions to explore are: How will AI affect the UK manufacturing sector in the future? Will productivity be enhanced leading to international competitiveness? Will there be a net displacement or creation of jobs? What skills will be required in the future?
The main indicators structured around two key measures of industrial transformation are: productivity or value-added per unit of labour employed; employment or jobs created and displaced. AI supplications such as expert systems and robotic process automation are pre-defined together with widely accepted definitions even when ambiguous and broad. The term automation is understood through boundary definitions and frames the theoretical linkages followed in the assessment (P. Nelson et al., 2023).
Conceptual Framework: AI, Automation, and Industrial Transformation
The formal definitions of artificial intelligence (AI), automation, machine learning, and robotics differ. Broadly speaking, an AI system makes decisions in an uncertain environment. Importantly, such decisions lead to automation, as they direct actions that then perform an automated task; yet the terminology underlines that automation can happen without AI. The automation of manufacturing tasks spans a long historical continuum, predating the first fulfilment of a formal AI capability.
Industry narratives suggest that manufacturing is especially vulnerable to the emergence of AI, because of its many routine tasks, yet the incident-based extraction of such narratives from general surveys suggests caution. Some sections of the literature suggests that the direct influence of AI on productivity, jobs, business modelling, and manufacturing business growth may be marginal, and that factors outside AI will dominate the evolution of manufacturing.
Could AI contribute disproportionately to productivity growth, labour-market outcomes, and the functioning of supply chains? Productivity remains fundamental to economic and social advancement, and the United Kingdom ranks behind most G7 economies. Several points of departure from pro-AI narratives warrant further empirical investigation. The UK does not lead G7 countries in the manufacturing adoption of robotics or wider automation, possibly the next best measure to AI. Thus, I suggest that AI and digital capabilities cannot expand upon current AI-enabled models for shop-floor task scheduling or inventory budgeting, unless those capabilities are first complemented by augmented or larger-scale automation. An initial loose aggregate parameter may exist, whereby the impact of AI upon productivity and business models is negligible without sufficient preceding automation.
Manufacturing comprises a key node for supply-chain resilience. Disruptions to supply chains induce widespread economic fallout, as the COVID-19 pandemic emphasised. AI technologies embrace several characteristics conducive to supply-chain management and resilience enhancement.
Current State of the UK Manufacturing Sector
Over the past four decades, the UK manufacturing sector has undergone significant transformations with noteworthy implications for its size, structure, outputs, and employment levels. Firstly, the overall share of manufacturing in value added has declined steadily, falling from 25% in 1980 to around 12% in 2020. Similarly, the share in employment has decreased from 30% to about 8% over the same period. These trends are mirrored in the broad Organisation for Economic Co-operation and Development (OECD) area, although the shares are considerably higher in comparison to other major economies, namely Germany, France, Italy, and Japan (P. Nelson et al., 2023). From another perspective, the UK stands out as one of the few economies that has also witnessed an absolute decline in manufacturing employment over the past four decades.
Manufacturing is dominated by a small number of highly productive sub-sectors, each accounting for 2–4% of total output and employment. The Food, Drink and Tobacco product sectors rank first, followed by Transport Equipment (eg. Aerospace, automotive, and rail) and Chemicals. The persistent significance of Food, Drink and Tobacco, despite its long-term decline, drives the noticeable contrast in productivity growth between the UK manufacturing sector and its overall economy and among diverse sub-sectors. Furthermore, manufacturing is distributed unevenly across the UK, with the South East, The Midlands, East of England, and the North West holding the largest shares in value added. The de-cumulative and relative declines add to the struggle in competing with flexible production in other countries across the value chain. The comparative international position on manufacturing—characterised by relative concentration and high shares in low-value-added activities is consistently weak.
Potential Impacts of Artificial Intelligence on Productivity, Employment, and Competitiveness
Adoption of AI technologies is predicted to boost productivity by 7–8 per cent per decade across the UK economy, with varied anticipated impacts on sub-sectors (ABRARDI et al., 2019). Current low levels of manufacturing AI adoption suggest substantial unrealised productivity gains, potentially exceeding 10 per cent per decade (P. Nelson et al., 2023). Learning curves exhibit a sustained positive trajectory despite performance converging to asymptotes. Many manufacturing processes, particularly in maintenance, supply chain, and quality control domains, remain amenable to substantial AI-enabled efficiency gains.
Policy measures that minimise investment costs and adoption uncertainties are crucial for realising these gains, especially for SMEs with limited cash reserves (Bughin, 2023). Initial investments are disproportionately larger than long-term expenditure, entailing a challenging upfront burden. Simplified safety standards, guidelines, and compliance frameworks can significantly reduce regulatory uncertainties surrounding AI system deployment and adaptive upgrades.
Productivity and Efficiency Gains
Artificial intelligence (AI) could enable UK companies to elevate their productivity by 30% or more, alongside application-specific gains (Chiacchio et al., 2018). The overall impact hinges on the pace of AI adoption: if uptake remains similar to non-AI automation, cumulative productivity improvements would reach approximately 2.1% by 2030; a gradual acceleration brings this to 3.9% (Harris & Moffat, 2019). Rapid integration over roughly the same timescale, although uncertain, could yield a 24% boost. These estimates substantially exceed anticipated advancements from alternative automating technologies. The likelihood of benefiting from new AI systems continues to rise, as reflected in longer average learning curves that favour manufacturers yet a significant proportion have yet to deploy any non-AI solutions.
For AI and non-AI capabilities alike, lack of suitable systems, high implementation costs, and resource constraints impede broader adoption. The need for complementary skills and training represents an additional bottleneck. Devised over several decades, the AI-enabled capabilities that offer supporting roles to manufacturers tend to build on existing, non-AI solutions.
Skills, Training, and Labour Market Implications
AI systems typically require fewer skills than previously required. Consequently, AI has the potential to reduce the industry’s skills level and increase the already unequal distribution of skills in the labour market. The introduction of AI technologies may negatively affect the skills of individuals in manufacturing occupations that already have low skills, namely, positions requiring lower levels of education (Chen et al., 2022). Work assistance technologies that provide assistance with decision making and problem solving tend to require higher levels of skills such as the ability to identify problems and develop solutions and critical thinking compared with standard workplace technologies such as word processors. AI usage is expected to increase the importance of collaboration with others, management of material resources, and operation and control, which are predominantly middle-skill tasks for occupations at the baseline.
Supply Chains and Resilience
Manufacturers increasingly adopt artificial intelligence (AI) to anticipate demand fluctuations, enhance inventory management, supply-chain disruptions, and supplier risk assessment. Higher AI penetration in manufacturing may consequently expand the scope for production-location relocation towards domestic markets and neighbouring territories, potentially stimulating regional diversification away from large UK agglomerations (P. Nelson et al., 2023) ; (Zapke, 2019).
Emerging AI applications in the UK manufacturing sector are reshaping operational choices and strategic priorities, with implications for productivity, employment, and national competitiveness.
Threats and Risks Posed by AI in Manufacturing
AI is intrinsically associated with various risks in the UK manufacturing sector, including strategic, financial, technological, and social vulnerabilities (P. Nelson et al., 2023). The first concern relates to the strategic and economic risks posed by high market concentration, increasing dependence on a handful of technology platforms to deploy AI solutions, and the need for national security precautions when AI is involved in critical infrastructure management. The second issue arises from technological or cybersecurity risks stemming from vulnerabilities in these AI-augmented systems; as manufacturers adopt AI tools to optimise operations, their reliance on data governance mechanisms and on third-party AI providers increases. The third risk pertains to social and regional disparities created by unequal access to the AI-enhanced productivity gains that other firms or sectors achieve. These gaps are evident in the stark differences in value added productivity between large and small firms in the UK, as well as between London and the North East, where projected gains from AI adoption may not be realised; consequently, there may be a need for proactive policies to foster such inclusive growth.
Strategic and Economic Risks
Several strategic and economic risks, indirect consequences of AI adoption in manufacturing, are particularly relevant in the UK context. One area of concern is the concentration of market power among a small number of platforms and providers. The rapid increase in manufacturing expenditure on AI-driven services recorded over the last five years raises questions about the long-term implications. The second risk is the increased dependence of businesses on a limited number of platforms and third-party suppliers. The use of AI is becoming intertwined with global supply and service chains, with potentially significant consequences. The third issue relates to national security, particularly in combination with platform dependence and information control. Some analysts have been describing the current geopolitical situation as a ‘new Cold War’, requiring a re-evaluation of British AI strategy, including AI in manufacturing. National security links to AI also impact research ethics, data sharing, and other aspects of AI development and deployment. Addressing these strategic and economic risks requires awareness and appropriate action, especially by the Government and the wider public sector.
Technological and Cybersecurity Risks
AI could introduce serious threats for manufacturing systems, emphasising the importance of comprehensive vulnerability assessments and risk prioritisation. Digital technologies create considerable new entry points for adversaries, as interconnected devices may suffer from software vulnerabilities and poor identity management. Interdependence of cyber and physical assets increases the risk of attacks on previously isolated operational technology (OT) devices, broadening the overall cyberattack surface. Compromised systems can produce operational downtime, equipment damage, or degraded product quality. Risk evaluation is critical to identify and prioritise security issues, enabling more effective protection and increasing resilience (Habibor Rahman et al., 2023).
Manufacturers universally adopt AI-driven tools and applications with little awareness of their information governance. Data constitutes a crucial production factor for a competitive AI-enabled ecosystem. Without adequate management, AI models could easily fall prey to alteration, theft, or unauthorised redistribution. Safeguarding data ownership and usage rights across the entire data lifecycle emerges as a major concern. Data governance challenges range from protecting sensitive information and upholding privacy, including compliance with the General Data Protection Regulation, to managing data provenance, safety, security, and ethical dissemination (Walz & Firth-Butterfield, 2019). AI solutions create dependence on providers of platform solutions or datasets, engendering further risks related to security, availability, and control.
Social and Regional Disparities
Digital technologies have already increased productivity and employment across the economy, but there remain starkly unequal, geographical, and socio-economic patterns in their uptake and diffusion. The average UK manufacturing firm has yet to benefit from a fifth of the AI efficiency dividends attainable within the sector, suggesting that economic model and data obtainment restrictions must be applied. Current automation risks are rebalancing towards smaller firms, with the activity share of high-automation processes diverging away from small-sized units. Small-scale manufacturers and those located in most deprived areas face greater socio-technical difficulties in applying AI technologies, partly due to concerns over upfront investment costs, training provision, and a lack of inhouse expertise. The introduction of wider adoption policies is indicated (P. Nelson et al., 2023).
Policy and Institutional Context in the United Kingdom
Current institutional context presents both opportunities and challenges for the adoption of AI in UK manufacturing. On the one hand, existing regulation and oversight regimes could hinder AI uptake. In the most recent regulatory report, the UK government highlighted the need to enhance the AI policy framework in respect to safety, governance, liability, copyright, privacy, and fraud. Although procedures governing digital technologies are in place, they trail behind frameworks for machinery or food safety, exposing manufacturers to legal and compliance risks. Indeed, unregulated AI could pose a greater threat than automation in industries already operating at high productivity levels, as illustrated in the automotive sector (P. Nelson et al., 2023). On the other hand, wide-ranging AI funding and support initiatives seek to encourage adoption and develop skills for the future of work.
Regulatory Frameworks and Standards
The UK’s AI Strategy promotes fast adoption of AI technology while implementing regulations to increase safety and mitigate risks associated with AI. The AI Safety Framework establishes key principles to guide the safe design, development, and use of AI, and the Data Protection and Digital Information Bill facilitates the control of AI training without restricting innovation. However, the balance between promoting innovation and establishing constraints is complex, as unnecessary delays can risk UK firms falling behind international competition. Furthermore, if harmonized standards are not ready when the AI Act takes effect, innovation will be further stifled during this period, as only companies with significant legal or compliance teams and capital will be able to innovate.
UK regulations impose minimal direct AI-related responsibilities on manufacturing firms. However, other pre-existing rules can become obstacles that slow AI adoption, such as the European market model which places an extensive array of rules and compliance burdens on companies relying on data and AI. The current and proposed institutional frameworks in the UK establish definitive principles for the deployment of AI and other advanced technologies in the manufacturing sector and indicate limited obvious risks, as hazards of direct misuse or concern over privacy remain low.
Innovation Policy and Funding
Artificial Intelligence (AI) has been the subject of much debate; there is, however, limited research concerning its consequences for the UK manufacturing sector. Such knowledge is vital for industrial, regional, and broader economic policy formulation (P. Nelson et al., 2023). This paper offers a policy-oriented assessment based on qualitative elaboration, targeting AI’s potential impact on productivity, employment, skills, supply chains, and international competitiveness in the manufacturing sector.
AI is recognised as an intrinsic threat from a strategic perspective. The UK has fallen to fifth in the global competitive attractiveness rankings; perceptions of the national security environment and AI-related supply chains have worsened; a significant share of mature manufacturers expect to cease operations in the next decade without capital upgrade and product transformation; and divergence in productivity growth rates has emerged, with the most and least productive firms progressing at markedly different speeds.
Skills and Education Policies
Between 1984 and 2022, the UK manufacturing workforce contracted 79% from approximately 5.9 million to under 1.25 million employees (Chen et al., 2022). The remaining manufacturing jobs are design-intensive and have a higher skill set. Regression analysis of Monthly Wages per Employment suggests AI could worsen wage inequalities across skill levels, and labour-market interventions are needed. Automated manufacturing falls under the Office of National Statistics (ONS) definition of routine-task intensity in jobs, which has implications on impact.
A McKinsey Global Institute report cites that only 25% of incumbents and two-thirds of AI-technology-leaders establish or reinforce fundamental reskilling and upskilling mechanisms while 30% of AI model-training-capable leading firms reallocate resources for corporate training signifying that even organisations aware of AI’s transformational potential often overlook human-centred approaches (Cacciolatti et al., 2017).
Strategic Scenarios for Mitigation and Harnessing AI in UK Manufacturing
1. The incremental adoption of AI technologies is likely in UK manufacturing: AI is expected to gradually augment existing products and services rather than gate a widely adopted transformation.
2. A sectoral-coherence scenario envisions coordinated and concurrent policies for modelling the AI and system architecture of mission-critical or productive manufacturing sectors, along preparation for the hi-tech and fore-runner sectors; both grammes combine under the remediation of the second stage.
3. Public-private partnerships for industrial policy address not only platforms but also collaborative entities or clusters comprising the national elements of data-exchange machine tools; policy frameworks focus on the selection of an appropriate multiplier for coverage or on the protection of the repository itself.
Incremental Adoption and Diffusion of AI Technologies
Incremental adoption and diffusion of AI technologies at the Enterprise level can take various forms to accommodate existing organization frameworks and pre-requisites. Enterprises can integrate AI based on:
(i) frequent determinations for predictive insight on business flow,
(ii) Ensuring process optimization through collaborative filtering,
(iii) Upgrading to build recommendation engine assistant, and,
(iv) Trading effectiveness at enterprise level through formulation and regulation of base production and AI automation for new trading format of service.
Parameters of enterprise requirement can be supported by customer focus and automation control considering long term growth and acceleration of productivity growth or new production channel／business mode (P. Nelson et al., 2023).
Productivity growth accompanies expansion of measurement and predictability, aiding enterprise substantially for expansion of production channel toward exposure of short term estimation toward reserved demand predictability (Bughin, 2023).
Sectoral Coherence and Regional Industrial Strategy
AI is defined as a general term covering a massive multi-disciplinary compound of techniques, including automation, robotics, machine vision, signal and image processing, expert and decision support systems, neural networks, deep learning, natural language processing, and general reasoning systems (P. Nelson et al., 2023). Consequently, the distinction between AI and automation is blurred in developing indicators of potential AI threat. In the context of productivity, AI influences efficiency through re-engineering business processes and reducing the search dimension of the solution space. The savings on these fronts are said to dwarf those possible through conventional automation. AI also has substantial system-level, firm-to-firm, and material disposition-wide impacts, referred to as value-chain effects. Importantly, not all automation incorporates AI, whereas AI can contribute to productivity improvements without triggering the installation of secondary automation equipment.
Public-Private Partnerships and Industrial Policy Instruments
The UK should consider creating more public-private partnerships to build industrial policy instruments that shape widespread adoption of AI. Such partnerships would augment investments already undertaken by the UK in this area, to make AI applications in manufacturing economically viable and to resolve critical issues related to social inclusion, skill mismatches, and environmental sustainability (P. Nelson et al., 2023).
Optimal characteristics of such partnerships in the UK would include financing by the public sector up to 50% of investment expenditures; active participation of mid-sized enterprises that are neglected by current funding programs, and governance through an open, broad, and impartial structure based on preliminary agendas set by a public authority as well as on significant industrial interest from all sides.
Evidence Gaps, Methodological Considerations, and Research Agendas
Data limitations constrain assessment of how artificial intelligence (AI) affects UK manufacturing productivity, employment, and competitiveness. Data gaps complicate measurement of AI adoption and output; macroeconomic measures do not capture sectoral adoption across broad categories; and most statistical data lack AI disaggregation. Finally, predicting the productivity effects of a multifactor technology like AI is intrinsically difficult.
Existing studies adopt diverse specifications: some rely solely on survey data, while others employ a wider range of quantitative and qualitative businesses (P. Nelson et al., 2023). One response is to formulate a mixed-methods research design comprising national surveys, ad hoc qualitative augmentations, targeted firm-level longitudinal analysis, and industry-level cross-sectional time-series measures. Priority datasets for UK manufacturing include the Office for National Statistics Annual Business Survey, the Organisation for Economic Co-operation and Development survey on AI in firms, and a UK-based survey by the European Commission on digital technology adoption, the Digital Economy and Society Index, the Business Register and Employment Survey, and the Inter-Departmental Business Register, alongside marker variables to a longitudinal data set for causal investigation (Upchurch, 2018).
Conclusion
The preceding analysis provides ample evidence that AI poses a considerable threat to the UK manufacturing sector. A cross-country benchmark analysis finds that the UK manufacturing sector is lagging behind other advanced economies in AI adoption. The empirical assessment also indicates that AI systems are likely to bring larger productivity and efficiency gains to UK manufacturers than non-AI technologies, coupled with competitiveness concerns stemming from slower adoption of non-AI automation technologies.
The economy-wide implications of AI adoption in manufacturing will perilously impact the UK’s productivity, competitiveness, and growth prospects in a context of high inflation, recession fears, global energy price shocks, and post-Brexit trading barriers. Similarly, a strong positive link between manufacturing productivity and capital investment in AI enables manufacturers to maintain competitiveness, instead of overly relying on rising production and operational costs to gain market competitiveness. Proactive investment in AI by manufacturing producers is important to achieve full recovery from immediate post-COVID-19 pandemic and COVID-19 disruption and sustain long-term growth, especially when the share of UK global total manufacturing value-added continues to shrink. To prevent further stagnation of manufacturing productivity, the focus on revitalising productivity in manufacturing and narrowing the post-pandemic productivity gap with other advanced economies should be complemented with investment in AI capital.
To mitigate the multifaceted threats posed by AI and to harness its benefits, the analysis highlights five strategic priorities. Secondly, the forthcoming AI Safety and Governance Framework should embed the challenges and opportunities presented by AI in manufacturing to channel additional funding and development such as the proposed Manufacturing Transformation Programme and the Recently-announced Industrial Technology Transformation Fund. Thirdly, dedicated policy measures should address urban-rural divides, firm-size disparities, and regional productivity gaps. Guidelines on Responsible Artificial Intelligence and the Strategic Framework for a Resilient Manufacturing Sector provide further opportunities to facilitate pro-competitive practices and cross-sector technology diffusion (P. Nelson et al., 2023).
References:
ABRARDI, L., CAMBINI, C., & RONDI, L., 2019. The economics of artificial intelligence : a survey.
Bughin, J., 2023. Does artificial intelligence kill employment growth: the missing link of corporate AI posture.
Cacciolatti, L., Lee Soo, H., Mar Molinero, C., Cacciolatti, L., Lee Soo, H., & Mar Molinero, C., 2017. Clashing Institutional Interests in Skills Between Government and Industry: An Analysis of Demand for Technical and Soft Skills of Graduates in the UK.
Chen, N., Li, Z., & Tang, B., 2022. Can digital skill protect against job displacement risk caused by artificial intelligence? Empirical evidence from 701 detailed occupations.
Chiacchio, F., Petropoulos, G., & Pichler, D., 2018. The impact of industrial robots on EU employment and wages: A local labour market approach. Bruegel Working Paper Issue 02/18 April 2018.
Habibor Rahman, M., Yazdandoost Hamedani, E., Son, Y. J., & Shafae, M., 2023. Graph-Theoretic Approach for Manufacturing Cybersecurity Risk Modeling and Assessment.
Harris, R. & Moffat, J., 2019. The decline of British manufacturing, 1973 - 2012 : the role of total factor productivity.
P. Nelson, J., B. Biddle, J., & Shapira, P., 2023. Applications and Societal Implications of Artificial Intelligence in Manufacturing: A Systematic Review.
Upchurch, M., 2018. Robots and AI at work: the prospects for singularity.
Walz, A. & Firth-Butterfield, K., 2019. IMPLEMENTING ETHICS INTO ARTIFICIAL INTELLIGENCE: A CONTRIBUTION, FROM A LEGAL PERSPECTIVE, TO THE DEVELOPMENT OF AN AI GOVERNANCE REGIME.
Zapke, M., 2019. Artificial intelligence in supply chains.