ARTIFICIAL INTELLIGENCE AND THE SMALL STATE
ARTIFICIAL INTELLIGENCE AND THE SMALL STATE
8th May 2025 – by Andrew Dolan
Unquestionably, the race to consolidate existing AI developments and develop even more ‘frontier’ capabilities is forging ahead. Perhaps it is no surprise that many of these new developments and applications are incubated in the world’s larger countries: natural resources, advantages in technical and industrial bases, financial muscle, educational attractiveness and sheer power have skewed AI development towards the major global powers such as the United States, China, the UK, France and a handful of similar-sized states.
Certainly, there are exceptions, such as Israel for example, a small state but a technical powerhouse or blocs of states such as the European Union, which seeks to shape AI development for the better of the whole and not the constituent parts.
Nevertheless, in terms of the ability to acquire huge amounts of data (especially to support and sustain AI training regimes), the ability to support moves towards Artificial General Intelligence (AGI) or even Artificial Super Intelligence (ASI) with vast arrays of computing power and the provision of well-endowed laboratories, much of it underpinned with high-quality educational establishments to facilitate algorithm research and development, larger states have numerous in-built advantages over small states. Add to this the ability to produce or procure the necessary chips and processing units or certain types of strategic minerals, which are required to develop AI, then this hegemony is unlikely to be challenged any time soon.
Therefore, as small states observe their larger counterparts dash to create new AI national strategies, invest and retool industrial and technical bases and collaborate with so-called ‘Big Tech’ giants, including investment in ‘strategic initiatives’, they must be wondering where they stand in terms of AI development? Can they ever be anything other than a downstream AI bystander, client or consumer?
As unattractive as this position might be, it is unlikely at this juncture that circumstances will improve or reshape small state agency in the AI domain. The factors that operate at scale and which help determine the future of AI capabilities or products remain, by and large, beyond small state national resources, whether this is access to vast data streams, computational power in cloud or quantum areas, or ‘cutting edge’ algorithm development. It is always possible that in terms of human resources, a small state can produce gifted AI engineers from out of its national educational and technical research and development reservoir but to exploit such potential usually means that the individual gravitates to the higher educational superpowers such as for example, Cambridge in the UK or MIT in the USA. If a small state cannot break out of its educational or technical limitations, the best and the brightest usually ‘vote with their feet’.
What’s left to do but wait and see where the next technological development will come from, try and understand what it can do and crucially, manage its eventual integration into society, whether this is business, public health, security or individual leisure?
It would be unhealthy for small states to adopt this wait and see approach and any measures that can be taken to appreciate and shape the nature of this AI accommodation would be well worth the effort.
Even now, small states should be studying the range of consequences that they might encounter as AI exploitation traverses through the everyday functioning of society. The speed and scale of AI penetration of society is staggering and therefore, preparing for either technical revolution or evolution will be time well spent. This means nationally understanding where and how AI applications, including the nexus with emerging technology, is likely to impact on state-level functions, the private sector and society. The elaboration of national strategies is no bad thing per se but it is vital for key public and private stakeholders to acknowledge that the fundamental factors underpinning such strategies, including risks and challenges, impacts on business and social behaviour, ranging from consumerism to investment can often become obsolete over a short space of time. Therefore, conceptual landscapes are less a fine art painting and more an impressionist sketch.
This regular shifting and modification of national responses to AI will also impact on those specific sectoral appreciations associated not only with National AI Strategies but perhaps more so on the specificities of discrete responses contained in Action Plans. It should be easier for small states to monitor individual business or societal responses to AI adoption and to generate genuine public-private frameworks for regulating how to garner the best from new AI applications and to deflect or prohibit the worst.
In fact, it might be more prudent for the authorities in small states to conceptualise and formulate new forms of AI National Strategies, with the essence being less about the direction and support in economic, business and finance or technical and educational sectors but rather on how best to coalesce around critical principles on how society remains cohesive in an environment where AI enhancements could be unevenly distributed in society and the country. In short, the new form of National AI Strategy could focus on how best to collaborate or partner with AI whilst retaining the fundamentals of what it means to be human.
To sustain such an approach, it again would be prudent to maintain a national AI capability in terms of higher education and training and technical research and development. It should be feasible within certain types of small states to rely on this sector to develop a corpus of AI coders and developers very familiar with algorithm developments, continuing to support researchers in trying to look for innovative solutions to the structuring of computational support mechanisms and even the production of safe synthetic data. Even if significant breakthroughs in AI development are likely to be limited from within small states, the creation of niche expertise even in a support function could help sustain national AI efforts.
Such a development might also be critical when it comes to choices regarding forms of AI products and its suitability for national use. There are unlikely to be few guarantees that ‘Big Tech’ companies will see value in creating too many niche or localised AI products, which might accommodate small state policy or cultural preferences but which lacks global or wider commercial attractiveness. In such circumstances, the ability to modify generic models to accommodate local sensitivities might be crucial.
The small state might also need to think about aligning itself to larger states or AI superpowers. Being a ‘client state’ is a familiar concept in the global geopolitical world, especially in the defence and security sphere. Having uninterrupted supplies of AI products or upgrades, including the AI form of cyber security ‘patches’ is likely to be increasingly moot, given that we are also witnessing the acceleration of moves in a global AI ‘Arms Race’.
For the small state, developing national defence champions in terms of military technology and weaponry is unlikely but it can and should invest in how its understanding of AI can help generate the foundations of a ‘societal security’ framework. This would be less an ‘Iron Dome’ missile defence system and rather an integrated and networked ability to protect critical network infrastructure and citizens from AI-enabled threats ranging from cyber threats to protection against manipulation of elections, media and external or internal threats to social cohesion.
What such a process might look like is far from fixed. Furthermore, the traditional tenets of international order are also far from stable. Perhaps, therefore, in terms of how best to manage the inevitable AI future, including the vanguard elements already here, the concept of National AI Strategies should evolve away from the traditional government approach to public policy and aim towards a wider use of SWOT Analysis templates when analysing risks and challenges and identifying utilisation and response frameworks for ever-more intelligent types of machine intelligence. Central to such SWOT analysis public policy framework will be a national capability to study and predict the likely directions and impact of AI developments as the most cost-effective way for small states to survive in the new AI ‘World Order’.