The Precipice of Artificial Hegemony
Share
As we mentioned in Machina Ex Diplomacy, AI has broken the bureaucratic barrier and has become a topic of considerable concern to operators within governments and intergovernmental organizations, who are now scrambling to identify the impacts and risks posed by the technology.
The spectrum of priorities is vast, with governments in the Global North focusing on the integration of the technology within their operational structures and keeping pace with the corporations leading the research in the field, while governments in the Global South are deeply concerned about the widening digital divide and are developing strategies to mitigate the technological lag in the race for AI.
On the intergovernmental level, the primary focus is on governance and regulation of the technology and ensuring that a modicum of transparency is kept when it comes to development, use, and research. One of the most pressing worries, reflected in legislative acts across the globe (like the EU AI act) and in UN resolutions is that the technological development outpaces the ability of governments and international mechanisms to govern and regulate its use.
The technology, with its elusive nature, its ability to navigate across borders, and its rapid pace of development, presents an entirely new set of challenges to cooperation on the global level. With a precariously balanced international order undergoing a state of transformation, and global and regional powers vying for positions in the new order, and with the technology itself presenting opportunities for competitive advantages, the choice for governments between cooperation and competition looms large.
The Race to the (Global) South
The digital divide between the Global South and the Global North – the difference in access to technology- is one of the most prominent topics discussed when it comes to international cooperation. Intergovernmental organizations, most prominently the UN, prioritize it as one of the key factors of achieving global development goals.
Governments however, while voicing support for the principle of shared and equitable access to technology and ensuring global development, tend to be reluctant to share technology with others without return. As actors that have their own self-interest as their primary goal, governments more often than not consider what they stand to gain from any given action, including providing support and assistance to others.
More powerful technologies exacerbate the principles of self-interst for governments; as they provide considerable advantages, it would be against the self-interest of any government to neutralize its own advantage in that respect. Take as an example the gatekeeping behavior exhibited by states that developed nuclear technology, and overlay it on a technology with the potential to subvert power structures from within the digital realm.
Bearing that in mind, the approach to cooperation from those who have access to the technology may be colored by the lens of opportunity rather than altruism; the growing demand for AI in the Global South may spur a race to be the first or among the first providers of the technology.
With the majority of countries in the Global South lacking access to capabilities and resources to develop their own AI frameworks at the same operating levels as corporations and governments from technologically advanced countries, and realizing that without access they will experience a widening digital gap, they have voiced their requests for support in the development of and access to the technology through multiple bilateral and multilateral frameworks.
This presents an unprecedented opportunity for corporations and governments that are actively developing advanced AI models. Unlike other technologies before it, AI, when integrated into operational systems is expected to have an impact on shaping the behavioral patterns of organizations that integrate it. It also ‘learns’ about the organizations, and can predict patterns and positions on a wide spectrum of topics and issues. Furthermore, through its integration with organizational structures, provides access to the internal workings of these organizations.
The opportunity in front of providers of the technology therefore, from a purely self-interested perspective, is that they will have access to and impact on the operational structures of the governments that they assist with this technology. Predictably, those offering access and support with the development of the technology cannot be expected to provide full access to the machinery behind its development and may be expected to instead provide limited versions over which they retain a degree of control and access.
Therefore, we can expect the race to be the main providers of support for AI for countries in the Global South to accelerate considerably both in the public and private sectors. Whichever actor takes precedence in incorporating their systems into the operational structures of recipient countries will have a significant advantage over those coming in at later stages; it is easier for bureaucracies to adopt an entirely new system than substitute existing ones.
Delivering the technology early allows providers to secure their presence within the recipient organizations. Systems in place within complex organizations (like government bureaucracies) are harder to replace once they have been incorporated. This is particularly true for systems like AI which integrate with existing infrastructures of an organization, and take time to customize and adapt to, and require investment of time and resources for training. This combination of elements results in entrenchment of the model within the structures of the organization, and triggers sunk costs considerations when changes are proposed, and may result in technological lock in once adopted.
While cognizant of these risks, governments of the Global South have a difficult choice to before them; either refuse to adopt systems created by other actors due to the associated risks, or to adopt the systems offered while hedging their bets to mitigate any negative impacts. Given the pace of development of AI, the risks of delayed adoption may be an insurmountable hurdle in terms of keeping pace with progress and having access to the technology, which will likely spur most governments to opt for the latter option, while investing in their own research.
On the governmental level, several global technological leaders have stated their intentions to support the development of AI in developing countries, and China has already taken an important initiative on the intergovernmental level in this instance. In cohosting the High-level Meeting on International Cooperation on Capacity-building of Artificial Intelligence along with Zambia on the sidelines of the UN General Assembly in September 2024, China effectively positioned itself as a focal point for the Global South on AI. During that meeting, it presented its AI Capacity-Building Action Plan for Good and for All, where it stated its readiness to engage with partners on capacity building, R&D, and developing AI language resources, and develop infrastructure, among others fields of cooperation.
The Global Zero Sum Game & AI
Governments across the world, more often than not, adopt an imperfect zero sum game approach to international relations. They assume that the gains of others are their losses, and they assume that they do not have all the information needed to make an informed decision. As a result of this approach, the global ecosystem is competitive and adversarial on everything from resources, to technology, to access to trade and navigation routes.
The multilateral system adopted in the wake of the second world war has made considerable strides in addressing this world view over the decades, bringing states together under the umbrella of the UN as a global platform for engagement and cooperation. However, this system is currently strained by the pressures of evolving power dynamics on the global arena. The growing mutual suspicion, increasingly adversarial positions, and the rise of conflict across different regions of the world are symptoms of a fraying multilateral system that is becoming increasingly sidelined by state and regional interest groups standing at odds with each other in the race for better positioning in the global hierarchy.
If precedent is anything to go by, governments will prioritize their own interests over any thoughts of collective good, essentially recreating a global prisoner’s dilemma. In this setting, it is theoretically in everyone’s interest to work for the collective wellbeing of the international community, and yet every actor, suspicious that every other actor is working in their own self-interest, accordingly works towards their own.
The risk of this approach when it comes to AI lies in the nature of the technology itself. Unlike other systems or preceding technologies, AI is developing in unpredictable directions and at an accelerating pace. It is also a technology that integrates with other systems of operation, and is being absorbed into existing systems on a global scale, becoming a more significant aspect of operational mechanisms across industries in the private and public sectors. It is also a technology that is open to customization, led by the private sector, and remains largely outside the bounds of regulation.
With all these elements in place, attempting to impose previous approaches to this unprecedented technology can spell disaster; individual interests vying to develop more powerful AI, and thinking primarily of besting competitors, can accelerate the race for powerful systems that deprioritize safety and security measures. The comparative focus on development versus guardrails is concerning; while there is extensive discussion of regulation in legislative and intergovernmental settings, the focus in development circles is on competitive advantage not restraint.
The problem with the competitive approach here is that the ultimate winner may end up being AI itself. With investment in ever more complex models capable of near human reasoning already on the horizon, we are a hair’s breadth away from technology that outpaces the human mind in reasoning. With corporations and researchers focused on the capacities of the technology, it stands upon the collective political will of governments worldwide to provide the balance.
While acknowledging the existence of an undetermined risk posed by AI throughout various legislation and resolutions, governments and intergovernmental organizations are approaching this technology in a manner far too bureaucratic to be effective. Establishing committees, meetings, and advisory bodies may have been an acceptable first step, but the current approach may find itself struggling to keep up with the pace of change and the potentials of the technology's impact.
The Pact for the Future, adopted by the UN General Assembly in September 2024, among other documents such as the aforementioned UN General Assembly Resolutions, reflect a global consensus on addressing the risks posed by AI. The recommendations of the AI Advisory Board established by the UN Secretary General call for various solutions such as establishing an independent scientific panel, a global fund for AI, and launching policy dialogues. The Advisory Body’s report further suggests enhancing cooperation on capacity building and establishing a data framework for data definitions and common standards, as well as establishing an office within the secretariat of the UN.
All these recommendations do not appear to reflect the urgency of the challenge. The timeframes proposed, annual or twice yearly intergovernmental meetings for example, are misaligned with the pace of development of the technology and its risks. Since the first resolution of the UNGA on AI in March 2024, AI has already undergone significant changes in capabilities, with the launch of new iterations of the technology with vastly superior capabilities than those of AI at the time of the resolution.
Conclusion
Given all these considerations, it would be beneficial for the collective good of humanity that governments come to an understanding of the magnitude of this challenge, and collectively figure out a way to set aside political and ideological differences when facing this issue.
The sculpture of politicians debating climate change (part of the Follow the Leaders Installations by Isaac Cordial) came to my mind as I studied the rate of development AI and compared it to the pace of governments' collective approaches to the technology. If we do not leverage the tools we have at our disposal to face this challenge before it outgrows out ability to deal with it, we may find ourselves floundering beneath the digital surface. If the digital divide between countries is wide, consider for a moment the digital divide that could form between humanity and a consciousness operating at ten or a hundred times the speed of human thought.
The complexity of setting up a mechanism that is focused on the collective good of humanity, devoid of political and ideological interests, that is geared solely on addressing the risks of AI reads like an idealistic pipe dream even to myself. Yet without prioritizing the collective good of humanity over the interests of individual actors in the face of what could potentially become the first non-human consciousness that poses a threat to our collective humanity, we may find ourselves facing a future described only in the most dystopic predictions of science fiction.
Read next
AI and Diplomacy: Addressing the Illusion of Expertise
The White House AI Memo: A Blueprint for Global Competition
