
Image generated by artificial intelligence.
The dream of creating an artificial intelligence or being has accompanied humanity for centuries. From Talos in Greek mythology to the mechanical knight imagined by Leonardo da Vinci, mankind has always sought to extend its capabilities and to reproduce certain mechanisms of its own intelligence. Artificial intelligence is therefore not truly a new technology. In its modern form, it has its origins in the 1950s, with the first work aimed at designing machines capable of reasoning, learning and solving problems
What is new, however, is the spectacular acceleration of its capabilities, its democratisation and its sudden arrival in almost every aspect of our daily lives. So much so that artificial intelligence has become one of the main power stakes of the 21st century. States, companies, international organisations and civil society are gradually becoming aware of the opportunities, but also of the economic, scientific, military and societal risks it represents. AI is indeed no innocuous technology. Its exponential capabilities appear poised to redraw tomorrow’s balance of power. How, then, can we ensure that this technological revolution continues to serve everyone?
AI models are becoming strategic assets
To date, there is no genuinely shared international framework for governing artificial intelligence. The major powers are therefore developing their own regulatory, technological and economic architectures. The United States favours innovation and competitiveness. China emphasises sovereignty and control. Europe is seeking to build a model founded on fundamental rights and regulation. Each defends an approach consistent with its history, its political culture and its strategic interests. It is natural for distinct world views to find expression. Artificial intelligence is, in many respects, a mirror of contemporary international relations. Reflected in it are our cooperation, our rivalries, our strategic ambitions and our visions of the future.
For several decades, for all its imperfections, globalisation carried the idea that it was possible to organise the world gradually around common rules. This movement was never linear or universal, but it had a direction. Today, we are witnessing the reverse. Geopolitical tensions are multiplying, sovereignties are asserting themselves and power politics is regaining strength.
We are gradually shifting from a world organised more around common rules to a more fragmented one, in which power relations are once again becoming the defining factor. This trend is becoming a little clearer every day in the field of artificial intelligence. Large models are no longer regarded as mere technology products; they are gradually becoming strategic assets. The recent debates over access to certain advanced models developed in the United States, the restrictions imposed on the most powerful semiconductors and the national security concerns that now surround certain technologies all illustrate this transformation.
One of the most unprecedented features of this technological revolution is that it is unfolding in a world where power is more dispersed than before. On the one hand, states are seeking to preserve their sovereignty and strategic interests. On the other, the leading models, computing infrastructure, data and research capabilities are concentrated in the hands of a limited number of private companies whose influence sometimes exceeds that of many states.
For the first time, the governance of such a transformative technology cannot be defined by a single power or by a single type of actor. This reality also raises the question of extraterritoriality. For several decades, the United States has demonstrated its ability to extend the application of some of its rules far beyond its borders.
Artificial intelligence could further reinforce this dynamic. Whoever controls the models, the critical infrastructure or the essential components also wields considerable normative influence over the rest of the world. The question, then, is no longer simply what rules will govern artificial intelligence, but also who will write them, in the name of what values and with which actors around the table.
The Global South increasingly asserts its determination to carry weight
In a world that has become both multipolar and technologically interdependent, the governance of AI cannot rest for long on the will of the major powers alone. It will have to bring together states, international organisations, technology companies, researchers, civil society and regions of the world that now refuse to be mere spectators of the rules that will shape their future. Yet the priorities, realities and aspirations of the Global South still too often remain on the margins of the discussions. And yet many of the economic, social and political consequences of artificial intelligence will be felt in these countries. The African, Asian and Latin American actors I meet regularly are not asking for governance that protects them. They are asking for governance in which they take part.
This distinction is crucial. For a long time, many countries in the South were regarded as recipients of rules drawn up elsewhere. That era now seems gradually to be drawing to a close. Without forming a homogeneous bloc, the Global South is increasingly asserting its determination to influence the major international balances. Whether on trade, climate, development finance, digital sovereignty or, now, artificial intelligence, the same demand keeps recurring: to be involved in setting the rules rather than discovering them once they have been adopted.
At a time when sovereignties are being reasserted all over the world, it seems hard to imagine that a governance model devised by a handful of major powers could durably be imposed on the rest of the international community. The more strategic artificial intelligence becomes, the stronger the demands for representation, participation and balance will be.
This is precisely where one of the main challenges of global AI governance lies: building a framework capable of fostering dialogue between the established powers, the technology players and the regions of the world that now refuse to be mere spectators of their own future. The question of legitimacy leads naturally to that of location. For the place where this governance is built is no institutional detail. It will shape the priorities on the agenda, the actors involved in the discussions, the compromises that will be possible and, ultimately, the rules that will govern one of the most decisive technologies of the 21st century and, ultimately, part of everyone’s future.
Since 2023, the international community has held a growing number of summits devoted to AI. From Bletchley Park to Seoul, from Paris to New Delhi, each gathering has helped to advance dialogue and bring more players to the table. But these conferences are by their very nature temporary. They encourage exchanges without necessarily creating the institutional continuity needed for genuine global governance to emerge. The model of travelling summits has opened up the discussion; it has not yet succeeded in building a permanent framework. That is precisely what explains the emergence of another contest, more discreet but just as important: the contest over where this governance might gradually take root.
Several cities and states now have strong arguments in their favour. Singapore has established itself as a credible player thanks to its regulatory expertise, its work on certification and its ability to bring Western and Asian technology worlds into dialogue. India is also gaining ground. Through its presidency of the G20, its initiatives in support of the “Voice of Global South” and its growing involvement in international digital debates, it has asserted itself as one of the leading voices of emerging countries on questions of technology governance.
The United Arab Emirates is investing heavily in artificial intelligence and seeking to position itself as a bridge between several regions of the world. Beijing and Shanghai boast considerable technological power as well as an industrial ecosystem without equal. Brussels already exerts significant normative influence through the European AI Act.
New York naturally enjoys the political legitimacy of the United Nations and remains one of the world’s leading centres of international decision-making. Each of these cities has genuine strengths. But each also carries the image, interests or priorities of the political system to which it belongs.
AI governance offers Geneva an opportunity
It is precisely here that Geneva stands out in a way that would be hard to replicate. For more than a century, it has established itself as one of the international system’s foremost venues for dialogue. More than forty international organisations, nearly one hundred and eighty diplomatic missions and institutions such as the UN, the WHO, the WTO, the ITU, the ICRC and CERN already exist side by side there.
Since 2017, it has also hosted AI for Good, which has become one of the world’s leading events devoted to artificial intelligence. In 2027, it will host the Global Summit on AI. But its chief asset is of a different kind: the trust it inspires. Over the decades, Geneva has developed a role that few places in the world are still able to fulfil: providing a space where actors with divergent interests agree to meet, talk and sometimes even forge compromises.
This ability rests on Swiss neutrality, the multilateral tradition and a culture of dialogue patiently built up over the course of history. In a world where rivalries are intensifying and bloc mentalities are gaining ground, this trust is far more than a diplomatic legacy. It could become one of the most valuable strategic assets for building global governance of artificial intelligence that is at once legitimate, inclusive and sustainable.
Even so, Geneva cannot rest on its laurels. If Switzerland aspires to play a central role in this emerging governance, it will also need to take a clear-eyed look at some of its weaknesses. The governance of artificial intelligence does not rest solely on principles, standards or regulations. It also rests on infrastructure: computing capacity, data centres, energy, digital networks, semiconductors and control over data. On this front, several countries have gained a significant lead. France, in particular, has made massive investments in its computing infrastructure, its data centres and its technology champions, aware that digital sovereignty has become one of the pillars of national sovereignty.
Yet Switzerland has considerable assets. EPFL, ETH Zurich, leading research centres, renowned international organisations, specialised institutes and a unique diplomatic credibility form a foundation that many countries might envy. The expertise exists. The talent exists. The networks exist too.
All across Geneva, researchers, international organisations, entrepreneurs, diplomats and independent institutes are already working to prepare for the major events ahead. Some are organising themselves, exchanging ideas and seeking to forge a shared vision capable of strengthening International Geneva’s value proposition in the field of AI governance. What they lack is not expertise. What they often lack is the ability to scale up, to bring more players on board and to receive support commensurate with what is at stake.
The question is all the more pressing because Geneva is at a turning point. The crisis of multilateralism, mounting geopolitical tensions and the financial disengagement of certain states, notably the United States, have weakened part of International Geneva’s ecosystem. Several organisations, agencies and NGOs have been forced to scale back their ambitions, with very real economic and human consequences for the region.
Artificial intelligence may offer Geneva a rare opportunity: not only to preserve its historic role as a venue for dialogue, but also to reinvent it around one of the defining issues of the 21st century. If the Confederation, the Canton and the City genuinely wish to make Geneva a lasting centre of global AI governance, they will need to do more to support the players already involved, back their initiatives, raise their international profile and invest in the infrastructure that will secure the credibility of this ambition in the long term. For a global capital of governance is not defined solely by the events it hosts. It is defined by the ideas it produces, the networks it drives, the talent it attracts and the solutions it offers the rest of the world.
Translated from French. The French original prevails.






