‘Sovereign AI’ is an emerging concept that focuses on national ownership of data and AI technologies. It emphasises creating localised, tailored solutions rather than adopting one-size-fits-all approaches. This shift has the potential to transform governments into innovation hubs and global AI leaders by leveraging unique datasets to address specific domestic needs.
Investing in sovereign AI could significantly boost GDP, with studies suggesting an increase of up to 15% by 2030. This growth is fuelled by AI’s capacity to revolutionise industries. Generative AI automates routine tasks, freeing up human capital for innovation and strategic planning. In healthcare, AI improves diagnostic accuracy, while in finance, it enhances fraud detection and reduces costs. A strong domestic AI industry also bolsters the broader high-tech sector, creating a skilled workforce and fostering global competitiveness.
As a general-purpose technology, AI has applications across the entire economy, driving productivity and growth. Its integration into personal devices, supported by advanced semiconductors, exemplifies its widespread utility. This adaptability makes AI a powerful driver of economy-wide transformation.
Sovereign AI also enhances economic independence and security by reducing reliance on foreign technologies. By fostering technological autonomy, nations can mitigate risks associated with external dependencies. Enhanced AI capabilities strengthen industries like agriculture, manufacturing, and logistics by streamlining operations, improving productivity, and reducing costs. These advancements create jobs, stimulate local economies and improve economic stability.
In the realm of national security, AI provides critical capabilities for threat detection, predictive analysis, and crisis management. Governments equipped with advanced AI technologies can better protect critical infrastructure, including energy, transportation and communication systems from potential disruptions. By enhancing situational awareness and resource allocation during emergencies, AI helps nations build resilience and ensure economic stability in an interconnected world.
Enormous potential, if done correctly
Globally, there are over 1,600 AI policies and strategies, with corporate investment reaching $67.9 billion in 2022, according to Deloitte. Leading players in the AI race include the European Union (EU), China, the United States and the United Kingdom, each adopting distinct approaches to development and regulation.
Several challenges must be addressed to realise AI’s full potential. One key issue is the establishment of regulatory frameworks to prevent misuse of technology and data, a process that is currently underway worldwide. Another challenge is overcoming automation bias, or what Erik Brynjolfsson calls the ‘Turing Trap’— the tendency to view AI solely as a replacement for human labour. This perspective fuels concerns about widespread job displacement, as reflected in media, business and policy discussions.
Policy gaps also hinder AI’s broader economic impact. To achieve widespread benefits, AI must be accessible to all sectors, including industries like healthcare, construction and hospitality, which often lags in adopting transformative technologies. While investments in technology and finance are significant, they alone cannot drive the economy-wide productivity gains needed. Policies must focus on accessibility, skills development and diffusion, particularly for small and medium enterprises, alongside retraining programs to prepare workers for AI-enabled job transformations.
AI offers a unique opportunity to address global economic challenges, including sluggish productivity growth, post-pandemic inflation and high borrowing costs. By relaxing supply-side constraints, AI can reverse productivity stagnation and potentially spark sustained economic growth. However, as Roy Amara’s law suggests, the impacts will take time—short-term effects may be overestimated, while the transformative long-term benefits are likely underestimated, potentially materialising meaningfully by the end of this decade.
Structural shifts in global supply chains, driven by disruptions like war, pandemics, and climate change, are also influencing AI’s role. Increasing focus on resilience and national security has fragmented traditional supply networks, moving them towards greater diversification, even at the expense of cost efficiency. Initiatives like reshoring and protectionist policies aim to mitigate risks, but they have also contributed to inflationary pressures. These shifts underscore the need for AI to support both economic resilience and sustainability amidst rising global complexities.
Secular trends
Beyond supply chain disruptions, longer-term structural trends are compounding economic challenges by limiting supply elasticity and increasing costs. Among these, declining productivity remains a significant concern, particularly in advanced economies. Additionally, aging populations, shrinking labour forces, and declining fertility rates are straining social security systems and creating fiscal pressures. Elevated interest rates have further complicated these issues, especially in countries like the United States and Germany, where labour shortages in high-demand sectors have contributed to slower growth and inflationary pressures.
The pandemic exacerbated sovereign debt levels across economies, with global debt now exceeding global GDP. The U.S. ratio stands at 120%, while Europe averages 88.6%, with some countries like Greece and Italy significantly higher. In contrast, China’s debt appears lower, though state-owned enterprises account for much of its corporate debt. Pandemic relief efforts successfully mitigated balance sheet damage, allowing demand to stay resilient despite rising interest rates. However, the fading of long-standing deflationary forces tied to emerging markets’ economic growth — described as the “Lewis turning point” — is also contributing to inflationary pressures.
A key concern is the long-term decline in productivity growth. For example, U.S. productivity dropped from 1.68% (1998–2007) to 0.38% (2010–2019). Sectors like tradable goods and services experienced a notable slowdown, while non-tradable services stagnated entirely. Despite this, the U.S. has outperformed other advanced economies, particularly Europe, due to better adoption of digital technologies and a more robust tech sector. Temporary productivity gains during the pandemic reflected a shift toward higher-productivity sectors and remote work, but it remains unclear if these improvements are sustainable.
The interplay between these forces has shifted global growth patterns from being demand-constrained to supply-constrained. This transition has resulted in subdued growth, persistent inflation, and elevated real interest rates. Structural conditions suggest borrowing costs will remain high, reshaping investment dynamics with higher capital costs and lower valuations. Although market expectations around rate cuts fluctuate, the broader economic environment points to a prolonged period of elevated rates.
These trends underscore the need for coordinated efforts to tackle productivity decline and foster sustainable, inclusive growth amid evolving economic headwinds.
Generative AI’s impact on traditional work
Generative AI represents a breakthrough in artificial intelligence, with a humanlike ability to operate across multiple domains and switch contexts based on conversational prompts. It can discuss topics like inflation, write computer code, and perform basic mathematics, though the latter remains an area for improvement. Its superhuman pattern recognition makes it a valuable digital assistant. Rather than full automation, the most effective use of generative AI lies in collaboration with humans—referred to as “augmentation.”
Generative AI is already having a profound impact on business and society. Its rapid expansion in capabilities and user-friendly nature makes it accessible to non-experts, allowing business users to harness it for tasks such as brainstorming, productivity enhancement, and new use cases. With more experience, users can refine their interactions to achieve greater efficiency, freeing up time for higher-value activities. As generative AI adoption grows, businesses can undergo transformative change — provided they address critical challenges related to Responsible AI, including risks to security, privacy, bias, and brand reputation.
The economic potential of generative AI is extraordinary, with the possibility of doubling global GDP growth within the next decade. It is poised to revolutionise white-collar work, but its benefits come with significant downsides. Developing economies that thrived on outsourcing, digitisation, and upskilling — such as the Philippines, India, and Kenya — face potential disruptions to labour markets and economic stability. No country is immune, and many are beginning to respond to the implications of AI on their sovereign interests.
The rapid adoption of AI, termed ‘flash growth’, has raised concerns about concentration of power. A small number of Big Tech companies dominate access to generative AI technologies, promoting a Silicon Valley-centric worldview. This centralisation puts critical decisions — such as how Large Language Models (LLMs) like ChatGPT or Gemini are developed — into the hands of a few. Additionally, national policies like export controls and tariffs could limit global access to AI systems, threatening economic and social development by fostering dependency on proprietary, ‘black-box’ systems.
Ethical concerns also loom large. Generative AI must be trained to understand acceptable behaviour, effectively giving Big Tech companies control over the societal and ethical values these systems embody. This raises critical questions about whose values are prioritised and the broader societal implications of these decisions.
The Capital Economics Study on AI’s economic impact
A recent study by Capital Economics reveals both the transformative potential of artificial intelligence (AI) and the uncertainties surrounding its global impact. Nearly 80% of surveyed clients believe AI will reshape the global economy. However, respondents were divided on whether this transformation will be primarily U.S.-centric, and over half expressed concerns about a potential AI bubble in financial markets, while a third remained undecided. The study emphasises that the global conversation about AI’s economic consequences has lacked a comprehensive framework for assessing its full impact.
Capital Economics positions AI as a general-purpose technology (GPT), comparable to transformative innovations like the steam engine, electricity, and ICT. Historical GPTs have reshaped economies in three stages. In the initial phase, adoption is slow, and productivity gains are minimal. For instance, during the Industrial Revolution, the U.K. experienced a 50-year period known as “Engel’s Pause,” marked by innovation but limited productivity growth. In the second phase, as costs decline and usage expands, productivity gains accelerate. Finally, diminishing marginal returns set in, and the pace of productivity improvements slows.
Past GPTs often took decades to deliver noticeable productivity benefits, though adoption lags have shortened over time. For example, steam and electricity boosted annual productivity growth in the U.K. and U.S. by 0.2-0.3%, while ICT raised U.S. productivity by 1.5% annually between 1995 and 2005. The faster realisation of gains from ICT was partly due to reduced delays in implementation.
The study highlights AI’s ability to boost productivity through efficiency gains, better workplace practices, and innovation. Efficiency improvements arise when AI either replaces human tasks or enhances workers’ productivity, allowing them to focus on more value-adding activities. For example, shared driverless vehicles could reduce parking space requirements, freeing up land for other uses.
AI is also expected to improve workplace practices by optimising processes like staffing based on predicted demand. Historical parallels include how electricity enabled factory layouts to move beyond centralised power sources and how digital communication facilitated outsourcing and offshoring. AI’s impact on productivity could be substantial, as shown by a 2023 study of call centres, where AI assistance increased agents’ productivity by 14%.
Moreover, AI’s role in innovation is potentially transformative. It could accelerate research by analysing vast amounts of data, identifying patterns, and suggesting new avenues of exploration. For instance, AI might aid material science by identifying chemical compounds to solve industry-specific problems, such as pollution absorption or corrosion resistance. This innovation-driven productivity boost could result in lasting growth, creating a virtuous cycle where AI continuously enhances itself.
Despite its promise, AI’s benefits may not be evenly distributed across sectors or countries. Historical GPTs achieved their greatest productivity gains when accompanied by social or political reforms that allowed full exploitation of the technology, such as the Enlightenment-era changes that supported the U.K.’s Industrial Revolution. Similarly, the impact of AI will depend on how effectively societies adapt to its capabilities.
Additionally, some sectors may see limited productivity gains from AI. For instance, automation in areas like targeted advertising and financial trading may only create firm-level efficiencies without improving overall economic productivity. Over time, as AI-intensive sectors thrive, less automatable, low-productivity sectors could grow as a share of GDP, eventually constraining overall growth.
AI’s impact on employment will vary by region and sector. According to a January 14 report by the IMF, 60% of jobs in advanced economies, 40% in emerging economies, and 26% in low-income countries could be affected by AI. While AI could enhance productivity in jobs requiring human oversight, it may also displace workers in tasks that are fully automatable.
The distribution of income will also shift, as capital income grows in AI-driven industries. This may widen income inequality depending on the market power of firms and the balance between capital and labour. Economists like Daron Acemoglu and Simon Johnson emphasise that AI’s deployment must complement human workers to support shared prosperity and offset job displacement.
AI’s potential to boost productivity hinges on its ability to spur innovation and address challenges like diminishing research productivity. Some economists, like Robert Gordon, argue that the scope for further innovation is limited, but others see AI as a critical tool to sustain progress. If AI accelerates innovation across sectors, it could drive sustained economic growth. However, the long-term impact will depend on whether AI complements or replaces human labour and how societies navigate the resulting economic and social shifts.
In conclusion, the Capital Economics study underscores AI’s transformative potential while highlighting significant uncertainties and challenges. By drawing on lessons from past GPTs, policymakers and businesses can better prepare for the profound changes AI is likely to bring to the global economy.