The combination of words 'anarchic world order' is quite paradoxical. Anarchy means lawlessness or chaos and order connotes a structured existence. Then it could mean individual or collective attempts to sabotage the order is prevalent. There is no other better nomenclature to articulate present state of international affairs. For around forty-five years present world order had seen off the cold war, almost a few decades of unipolarity. Current state of international affairs can arguably be termed as multipolar. In a multipolar world, when there is no balance of power equations are working, instability is set to creep in thereby disturbing international peace and security. Google Gemini quoting the UN tells that at present more than 130-armed conflicts are decorating the international scenario in which state participation is the highest since the Second World War. In such a situation, it is pertinent to examine the dynamics of world order as we have been familiar with since the inception of the UN.
Triangular Relation among IR, International Order/Law and International Organizations
International Relations
International Organaisation International Order/Law
International relations whether in the form of bilateral or trilateral or quadrilateral or regional or multilateral conducts on the basis of pacta sunt servanda shape international order or Law. For example, when nations decided to participate in the Uruguay round of trade negotiations seeds were sown for a multilateral rule based trading system. This international order in return structures state behavior too. Countries are cajoled or coerced to follow the multilateral trading rules. On the other hand regional or multilateral relationships are the basis for the formation of international organizations/platforms. Most of public international law is formulated under the auspices IOs. It is true that Uruguay round culminated in the WTO and the covered agreements. But post-WTO development of international trade law through Ministerial Conference decisions and Dispute Settlement Body decisions are of equal importance. IOs determine how nations carry on themselves in the international plane. In other words there is no denying of the fact that IOs force international relations to adjust while being the umbrella structures for initiating and concluding multilateral agreements. Having articulated this much there is no need to ponder over the relations between IOs and international order/law. These triangular pushes and pulls ae the backbone of the international order since post second world war. Most contributions were multilateral in flavor. That phenomenon is undergoing a transformation now. We are witnessing more bilateral/quadrilateral type attempts now days. This part can be summed up with a comment of Dr Jayasankar, Indian Minister of External Affairs in the India-Arab Foreign Ministers' Meeting in 2026. He opined that cross-border terrorism is unacceptable as per the fundamental principles of international relations and diplomacy. He was not even bothered to cite the UN Charter or international law. Does this signal a retreat from the civilization structures international community painstakingly settled for after the world war? Is this the end of globalism as we know?
Challenges Troubling Present World Order
No better affirmation is needed for the perils of a multipolar world order than what Dr Sasi Tharoor wrote on the editorial pages of the Hindu (May 30,2026) on whether international law is optional for powerful countries. Still, one needs more clarity on the canvas. Remember the journey was started in 1950 through the Uniting for Peace Resolution when the collective strength of the General Assembly assumed secondary responsibility to maintain international peace and security in the void created by the cold war powered Security Council veto. A gradual shift in the attitude of global community culminated in the Iraqi invasion in 2003 what late Prof VS Mani in one of our weekly JNU seminars called the notorious “authorization of contract killing by the Security Council”. Rest is history. Today an Organization and its functionaries which promised to save 'successive generations from the scourges of war' is a silent spectator in the face of scaling violence around the world. UN Charter principles are violated with impunity. On the other hand, demands for restructuring of the Security Council in line with emerging realities in international politics have remained inattentive for many decades.
It is not the UN system alone even while accepting the fact that several non-celebrated international agencies like International Telecommunication Union, International Postal Union, WIPO etc are doing their routine jobs silently making international community life possible. The crisis is one of normative erasure and gradual fading of organizational relevance. During COVID period the Trump administration had given the letter of intent to withdraw from the commitments in relation to the World Health Organization. This put the future of global public health in jeopardy as the US had been the major budget contributor to the WHO. Google Gemini from White house records intimates that the US administration has decided to withdraw from 66 international platforms including the Intergovernmental Panel on Climate Change and UNCTAD. The future of Paris Accord without the US participation is in tantrums.
Implementation of international norms was never a major strength of IOs. But the weight international public opinion carried in is missing in the present times. Reporting procedures are ritualistic whether in the human rights or environmental parlance given the importance given to comparative trade advantage by authoritarian look like governments. Though implementation of international norms was never strength, recourse to justice or in other words a functioning dispute settlement process is a badge evidencing a system in action. The International Court of Justice preliminary ruling on Rohingya genocide was not given proper attention despite Myanmar Junta regime is being relatively low profile, let alone the fate of advisory opinions. Dr Sasi Tharoor mentions about failure of the UN Law of the Sea Convention. ITLOS, after three decades has only a jurisprudential asset of thirty odd rulings. The contempt showed to the Permanent Court of Arbitration ruling in the South China sea case is a part of Chinese folklore by now. Warrants form the International Criminal Court to nation-heads are thrown to the dustbin. The US taking hostage of the WTO Appellate Body mechanism on the face of it has not affected the volume of trade under the WTO framework much. Gemini tells that more than seventy percent of the world trade happens under WTO rules. But if the stalemate continues, the WTO would be moving gradually towards an organizational irrelevance. It is unclear whether international attitude shapes national approaches or vice versa. Anyway, the minimum normative standards developed by international community over seventy years or so are in calamitous condition. The situation that any country can do anything with impunity might be a feature of the prevailing multipolarity. Our concern here is how AI global race has positioned itself in this political multipolarity.
Bipolarity in AI Global Race
There are two economic and military superpowers in the international scene with China being the not so distant and a little bit hesitant second. A host of middle power countries make up the multipolar world. In the global AI sphere too, a replica maybe found with small variations. Kai-Fu Lee (2018) detected 'a new kind of bipolar order' in this respect with China and the US miles ahead in the AI competition. UK, France and Canada don't have the 'venture capital ecosystem or larger user bases' to generate data. He predicted an AI revolution with four waves with Internet AI being the first one. In this category AI algorithms are used as recommendation engines which make use of labeled (linking a piece of data with a specific outcome like e purchase and consumer habits algorithm to weed out fake news etc.) digital data. The US and China stood on an equal footing in this wave at the time of prediction, he predicted a 60-40 advantage for China in the future because of the data leverage. Business AI is the second wave. According to Kai these AI 'mines databases for hidden correlations that often escape the naked human eyes and brain'. But he keeps a silence on the matter that whether these correlations are logically or socially plausible. With its early adoption of business AI in financial sector, the US enjoys a 90-10 advantage over China. China is expected to narrow down this to 70-30 in five years. Perception AI is the third wave in AI revolution. Even in 2018 itself there was a proliferation of sensors and smart devices which 'turn on physical world into digital data that can be analyzed and optimized by deep learning algorithms' which have great potential in commercials, education and hospital sectors, teaching etc. China depends on domestic stat ups for the AI hardware which gave a 60-40 advantage over the US which depend on global production chains. For the same reasons the advantage may shoot up to 80-20. Autonomous AI is the last in the AI revolution. It may be that the world was not familiar with agentic AIs in 2018. This wave rides on the 'integration and culmination' of first three waves. It combines machine's ability to optimize from extremely complex data sets with sensory powers which exponentially expand tasks like power of sight, sense of touch and ability of optimizing from data). Autonomous robots, self-driving cars and autonomous drones are the examples. In self-driving cars the US has an advantage but with the right policy decisions China is catching up. Economic advantage turned technological progress alone are not just crucial in the AI revolutions. It is the open-mindedness to innovate and a care-free attitude to risks which matter. Only time can tell which one move in the right direction, whether the care-free attitude or a more cautionary European attitude.
Theoretical Challenges to Global AI Governance
Those who see threat in AI development because of risks[1] like labour, copy right, environmental, potential of AI becoming a sentient being etc. are arguing vociferously for a global regulatory regime ensuring safety. Neil Turkwitz (quoted in Gary Marcus,2024) commented that “ without legal accountability for AI harms, all the architecture and self-regulation is merely compliance theatre”. Emily and Alex express zero trust with self-regulation as they feel that regulators should not assume that companies can actively self-regulate. Proper channelization of innovation need regulations to shape it. The expertise needed in regulatory matters is not technological but expertise in “socio-technical systems and other civil rights”. There are several who suggest techno legal solutions. According to Arvind Narayanan and Syash Kapoor (2024) AI safety community believes that technical solutions aligning AI with humanity's interests will be found and there should be a ban of powerful AI being built at all. They consider both these as non-feasible because it is still unknown how many rungs are remaining in the ladder to achieve generality and each step needs a scientific breakthrough. Every new AI is fundamentally different from that is in the just preceding rung. Safety alignment with AI research to a hypothetical future is inherently limited. Both occur concurrently. They feel that AI safety community is too optimistic about regulations which would require draconian surveillance of data centres and computing resources (of disproportionate use) which might need unprecedented international cooperation. Emily and Alex (2025) feels that the demand for 'brand new legislations' by both boomers and doomers is a delaying tactic by AI companies and lobbyists to any set of limitations on their technological and business projects. They claim that 'a good deal of regulatory tools' are there to deal with exaggerated corporate claims like monopolistic laws, consumer laws, civil and labour rights law to extend to cover ill effects of AI. Gary Marcus(2024) lays down five reasons for the possible conclusion of an international governance of AI 1) Nations do not want to abdicate their authority to technology companies when there is chance to assert it;2) to prevent cyber criminals from using new technologies to manipulate markets and citizens at an unprecedented level ;3) to prevent deceleration of the climate catastrophes by addressing the environmental costs of AI ;4) to prevent surrendering to superhuman intelligence devoid of human values and prevent conducts like forum shopping (he suggests something like International Civil Aviation Organisation and International Atomic Energy Agency). To Gary Marcus “we can't expect global AI governance to work until we get national AI governance to work first”. Both national and international law shall work together. More than the techno-legal challenges the uphill task could be a proper meeting of all nations' minds which is difficult to attain in a multipolar world.
International Initiatives
Within these differences, international community has made certain progress after understanding the risks associated with AI development. The OECD formulated AI Principles in 2019. G20 too in the same year formulated some AI Principles. Council of Europe (2024) delivered the first binding Framework Convention on AI which casts a duty to respect human rights, democracy and law. UNESCO's Recommendation on the Ethics of AI suggests responsible AI with ethical standards respecting human rights and the environment. G7 initiated the Hiroshima Process on Generative AI agreed on a comprehensive policy which mainly stand for the free cross border data flow. The UN Secretary General formed a high-level AI Advisory Committee(2023) which was the backbone behind the UNGA Resolution (non-binding) in 2024. The Resolution calls for a human centric, reliable, explainable, ethical and inclusive AI development. It embarks on risk identification, evaluation, testing, risk prevention and mitigation and data governance. It recognize the importance of intellectual property rights and recommends human oversight of AI tools and development to ensure transparency, predictability, reliability and understandability. Several other platforms like the WIPO, ITU, ISO/IEC etc are dealing with the with the AI matter.
A Binding Global AI Regime and the Probabilities
Considering the present multipolar turbulence, nothing definitively can be said about negotiations for binding AI regime at the international level. We can only guess some probabilities. One probability is that there will be no international regime for AI since the US and China have vested interests to boost their industry in line with their domestic laws and policies. Europe with a stringent AI governance model will have to align the rest of the world who majorly happen to be consumers of the technology to counter the AI hegemonic forces. The second probability points towards a weak declaration type instrument even modifying the UNGA Resolution standards to suit the US and China needs. Another probability is the Framework-Protocol approach giving everyone the breathing space to get ready for a binding Protocol in the line of UNFCC. The last probability will be a watered down binding instrument to suit industry needs because regulation itself is functionalism. It gives legitimacy to something which is a taboo. Anyway the greatest challenge would be to bring all present day selfish actors to the negotiating table.
Views are personal.
Bibliography
Kai-Fu Lee, AI Super Powers, China, Silicon Valley and the New World Order (Harper Collins,Boston 2018)
Emily Bander and Alex Hanna, The AI Con, (Penguin Random House,UK,2025)
Gary Marcus, Taming the Silicon Valley, (MIT Press,London,2024)
Arvind Narayanan and Syash Kapoor, AI Snake oil, (Princeton UP,New Jersey,2024)
Some AI companies demand permission to do AI research only for licensed companies which is a ploy to stifle competition and transparency in the discussions on technology development. Arvind and Sayash predict more threat from AI misuse than from the technology itself. ↑