Artificial Intelligence and Middle Powers : Navigating Sovereignty, Opportunity and Risk

EDAM, 22 Kasım 2024

In today’s rapidly evolving technological landscape, countries are grappling with the opportunities and risks presented by artificial intelligence (AI). For middle powers, this creates a particularly unique challenge. Leading powers such as the U.S. and China command vast resources and can control their destiny, at least to some extent. Smaller, more nimble countries such as Singapore and New Zealand can pivot quickly with boutique solutions. However, the middle powers need to balance multiple considerations concurrently – national sovereignty and globalization, AI’s benefits and risks, personal liberties, and regulation.

The Middle Power Challenge

Türkiye is a case in point.

On the one hand, the country has the largest population in Europe – the 22 million people under the age of 18 is larger than the populations of all but six EU member states. Türkiye has been among the top 20 global economies for decades. Turks are proud with their imperial heritage, the War of Independence after the First World War, and strong military, which is the second largest in NATO. 

On the other hand, Türkiye represents about just 1% of the global population, economy, and trade. Its application process to EU membership has stalled. The high technology represents just 3% of Turkey’s total exports.

Türkiye needs to preserve its sovereignty, reap the benefits of AI, and protect the personal rights and liberties of its citizens – all at once. This requires striking the right balance in governance, performance, and regulation. The Swiss alchemist Paracelsus’s words come to mind: “The dose makes the poison.” 

For middle powers like Türkiye, the challenge lies in finding the right dose.

 

Preserving the Sovereignty

Historically, sovereignty has been about writing and enforcing the rules —the codes— that govern societies. In our age, there are new players who write codes – this time the computer codes and associated algorithms that govern how we shop, get our news or socialize with other people. AI adds further complexity to this equation by blurring the authorship, and hence responsibility, of the code.

As historian Yuval Noah Harari observes: “The ability to hack humans means the power to manipulate and control them. And AI will be the key technology of the 21st century. Whoever controls this technology will dominate the world.”

We are looking at multiple layers of sovereignty, each with its own challenges:

• Global – International organizations or multinational companies (the Big Tech) have the reach and potentially the control over the developments in AI but do not have the mandate to govern and legislate
• National – nation-states, particularly the democratic ones, have the mandate to govern and legislate but lack the global reach and the control over the developments of algorithms.
• Individual – At the personal level, free will is dramatically challenged by algorithms and AI’s omnipresence, putting the very concept of a sovereign citizen in question.

To navigate this, middle powers may redefine sovereignty as a “trust architecture” and act accordingly to achieve the optimal solution for their citizens.

This architecture needs to have three main characteristics. First, it needs to be dynamic, given the pace of technological change – one can never say they have ‘found the solution’. 

Second, it needs to be holistic, given the wide range of activities AI affects – security, economy, science, arts and culture, among others. 

Third, it needs to bring as many as stakeholders as possible, likely under various overlapping structures. This might mean ‘a regime complex that is multi-multilateral, comprising several institutions and initiatives, each involving different membership groups’, as defined by Emma Klein and Stewart Patrick in their Envisioning a Global Regime Complex to Govern Artificial Intelligence paper published by Carnegie Endowment. 

Under such a system, ‘many institutions for AI governance will be intergovernmental, with membership restricted to sovereign states; some will have universal membership, whereas some will be narrower, selective, mini-lateral frameworks among like-minded nations. Other global arrangements will have multiple stakeholders, involving not only national governments but also corporations and civil society actors. Eventually, some normative commitments may become grounded in binding international law, while others will remain voluntary’, as noted in the mentioned article.

Unless they play a role in establishing the new governance framework, middle powers risk either becoming vassal states of large countries/ Big Tech or being reduced to irrelevant backwaters with ‘left behind’ populations.

Reaping the Benefits

For middle powers, AI offers tremendous opportunities for political, diplomatic, national security, economic, and social advancement. AI should be seen not as a technological tool but as the major driver of a development effort based on leapfrogging rather than chasing developed countries.

• Political – AI could make government affairs more efficient, effective, and democratic.  While this is valid for all countries, it is particularly relevant for middle powers, many of which suffer from red tape, suboptimal decision structures and retreating democracies. Numerous bureaucratic tasks could be handled by personal digital government officers running on AI, saving time and financial resources. Decision-making could be supported by AI – preparing questions for college entrance exams, forecasting inflation daily based on big data analysis and setting benchmark interest rates could all benefit from this emerging technology. It might not be surprising to see a Siri-like minister in cabinets of coming decades. The democratic scope could also be expanded given that the suggestions, complaints and preferences of citizens may now be collected, analysed and synthesized by AI much more easily. 
• Diplomatic– AI could help middle powers expand both their soft and hard power. On the soft power side, helping shape the global conversation around responsible AI development at existing (UN, G20, OECD) or new forums and exporting AI governance models, ethical frameworks, or regulatory standards could be possible. The potential migration from bi-polar world to a multi-multilateral system could provide opportunities to middle powers that have capabilities to contribute to the global discourse. The Global Partnership on AI (GPAI), led by nations like France and Canada, and the UK’s efforts to be a leader in AI Safety, as demonstrated by the Bletchley Summit, are examples of this approach.
• National security – Using AI in predictive analytics for threat assessment, cybersecurity, counterterrorism, and military applications in the field would improve security capabilities. Unlike nuclear weapons, AI-based defense capabilities could be developed without the vast resources of superpowers. This ‘democratization’ also creates a potential risk for the international system, as demonstrated by the cyber activities of rouge states such as North Korea or various non-state actors.
• Economic – Middle powers, many of which missed out on the Industrial Revolution, could leapfrog developed countries by investing heavily in AI. They could position themselves as regional hubs by leveraging their populations to train AI models, maintaining relationships with industry players from multiple spheres of influence and introducing smart regulations, such as sandboxes and test areas.

AI could be the linchpin in productivity, rather than resource, driven development. Take the case of Türkiye. While the real GDP has grown by almost 5% per annum over the past four decades, only 1% of this was due to factor productivity growth. As a result, Türkiye has the 18th largest GDP, but 72nd largest GDP per capita, on par with the global average.

• Social – Education is perhaps the most exciting frontier for AI in middle powers. AI has the potential to democratize access to information and provide personalized training to individuals. This is a fascinating opportunity (as well as a threat) for middle powers, most of which have relatively young populations. That said, this also calls for a broad-based education reform, redefining the roles and capabilities of teachers. For example, Estonia, is already incorporating AI and digital literacy into its education system, demonstrating how nations with relatively limited resources can lead in this area.

Success in all these areas depends on a nation’s AI readiness. The IMF’s AI Preparedness Index shows a stark contrast between advanced economies (index score: 0.68), emerging market economies (0.46), and low-income countries (0.32). Türkiye ranks 50th out of 174 countries, with a score of 0.54, reflecting gaps in digital infrastructure, human capital, technological innovation, and legal frameworks. 

Unless middle powers overcome these challenges, national security could be compromised, existing global income/ wealth disparities might be augmented, many jobs could be lost, social fabric could be splintered, and these countries may never be able to converge advanced countries. 

The society is aware of this challenge – a global survey by Ipsos shows that while 71% believe AI can solve problems, 57% fear it is destroying lives!

Protecting the Rights

Let’s consider some of the key underlying principles of liberal democracy: Free will, individuals’ right to make their own choices, consent of the governed, tolerance of differences, freedom of speech and press, equality before the law, rational decisions.

All of these concepts are challenged, if not are under outright attack, by social media, algorithms and AI. Are we really making our own decisions if we are continuously ‘nudged’ (if not manipulated) by algorithms? Can we really talk about a free public square if we are confined to our own echo chambers? Whom do we consent to govern us, if the ‘codes’ are determined in the remote headquarters of for-profit companies or even by nameless AI robots?

The middle powers need to safeguard their citizens against three key challenges:

• Techno-autocracy. ‘All seeing’ monitoring systems, coupled with analytical capabilities of AI could pave the way for a dystopia. Citizens could always be under surveillance; rated, rewarded and penalized based on certain criteria; and effectively reduced to minions. We shall reject this road to tyranny and serfdom.
• Big Tech dominance. Large technology companies are not white knights of liberty or innovation, as they are sometimes portrayed in middle power countries with less than perfect democratic credentials. Their main purpose is to maximize their shareholder value as aggressively as possible. Institutions with responsibility and mandate to defend public good, such as parliaments, need to engage with them in a constructive manner, but always keeping in mind different objective functions in various topics including innovation, competition or rights and liberties.
• Anarchy. In the world of disinformation/ misinformation, identity theft, deep fakes and infringement of copyright by AI models, a ‘wild west’ approach is not appropriate. That said, potential overregulation also risks stifling innovation in the name of safety. We are back to Paracelsus’ warning about the right dose, which would change quite dynamically.

 

The Way Forward 

As Harvard professor John Kotter reflected: Managers deal with complexity, leaders deal with change. In the age of AI, we need to cope with both complexity and change. Middle powers must also create a solution space that operates across multiple layers:

• • Global – Leverage the emerging multi-multilateral regime complex to contribute to global problem solving and help shape the emerging discourse.

• National – Put AI to the center of the ‘leap forward’ strategy in political, diplomatic/ national security, economic and social development program. Establish smart regulations to guard democracy, rights and liberties. Develop state capacity given that leaders in government, parliament and bureaucracy will need to manage unprecedented challenges and negotiate with extremely smart, well-compensated and aggressive counterparties in superpowers, Big Tech or rouge actors. Remember words of Geoffrey Hinton, one of the Nobel Physics Prize laureates of 2024, who quit his role at Google to speak more freely about his growing fears about the risks of AI to humanity: ‘’We have no experience of what it’s like to have things smarter than us.”

• Individual – Support and equip citizens to adapt to the AI-driven world so that they do not become losers of this groundbreaking change or techno-fatalists passively watching developments from sidelines. As futurist Alvin Toffler once said: “The illiterate of the future will be those who cannot learn, unlearn, and relearn.”

The future belongs not to the largest or smallest players, but to those who can adapt. Middle powers such as Türkiye should leverage the AI disruption for a development leap. With clear strategic thinking, high-quality state capacity and well-balanced approach, I am confident that this can be accomplished.