Global AI Governance Cooperation Needs to Carry on Despite Transatlantic Disputes

Conventional wisdom suggests that these days the only really big fight around AI takes place between the US and China and the rest of us are sort of idle observers of this epic struggle. Furthermore, geopolitical logic seems to imply that the West at large will inevitably go with the US, while a very significant part of the rest will stick to China.

However, this simplistic black-and-white picture might be misleading. Just look at the outcomes of the landmark AI Action Summit that took place last month in French capital Paris. The outcome of the Paris Summit suggests that the real line of demarcation in the AI universe runs not between the US and China, but rather between the Anglo-Saxon world and all the other international actors, including the European Union. In the vast AI space, clashes of interest between Washington and Brussels turn out to be no less significant than disagreements between Washington and Beijing.

In Paris the European Union managed to find a mutually acceptable compromise with the Global South, India turned out to be able to overcome its frictions with China, and only the US and the UK preferred to challenge the emerging global AI consensus.

It is yet to be seen how the ongoing transatlantic disputes will be resolved. It is clear, however, that the disagreements go far beyond the AI domain, reflecting much more profound political, social and even civilizational divergences between the Anglo-Saxon and the Continental European worlds. Nonetheless, the numerous AI management challenges cannot wait until both sides of the Atlantic Ocean sort out their disagreements and come to a common denominator. The AI global governance should move on—with the US and UK governmental participation, or even without it for the time being.

Conventional wisdom suggests that these days the only really big fight around AI takes place between the US and China and the rest of us are sort of idle observers of this epic struggle. It argues that we can cheer on Washington or Beijing, while enjoying many fascinating features of GPT-4 or of DeepSeek, but the main line of demarcation in the emerging AI space has already been drawn and this space will inevitably be divided between a bunch of omnipotent US and Chinese high-tech giants. Furthermore, geopolitical logic seems to imply that the West at large will inevitably go with the US, while a very significant part of the rest will stick to China.

Anton Kolonin:
Growth Paths for the Coming Superintelligence

However, this simplistic black-and-white picture might be misleading. Just look at the outcomes of the landmark AI Action Summit that took place last month in French capital Paris. Sixty-one participating nations, including France, China, India, Japan, Australia and Canada, signed the final declaration stating the goal of keeping AI “open, inclusive, transparent, ethical, safe, secure and trustworthy, taking into account international frameworks for all" and "making AI sustainable for people and the planet.”

However, two major countries—the US and the UK—refused to subscribe to the document thus raising doubts about whether a global consensus on developing and regulating AI technologies was attainable. The US and the UK clearly prefer flexible AI guidelines to any binding regulations, with the noteworthy exception of the restrictions on hi-tech cooperation with China.

The outcome of the Paris Summit suggests that the real line of demarcation in the AI universe runs not between the US and China, but rather between the Anglo-Saxon world and all the other international actors, including the European Union. In the vast AI space, clashes of interest between Washington and Brussels turn out to be no less significant than disagreements between Washington and Beijing.

Anatasia Tolstukhina:
US Technology Policy amid Rivalry with China

Most European nations are less confident that market mechanisms on their own can provide magic solutions to the numerous problems of AI development. The EU economic model implies more rigid regulatory frameworks and more attention to the potential social repercussions of high-tech development. Therefore, the existing EU legislation mandates strict rules for high-risk AI applications, focusing on safety, transparency and accountability.

On the other hand, many EU member states are reluctant to break their long-standing partnerships with China, including AI cooperation. The US and UK demands on fully terminating these partnerships are often perceived in Europe as excessive, inappropriate and explicitly self-serving, therefore, it comes as no surprise that the line of demarcation in Paris questioned the perceived Western cohesion in the AI area.

This is not to say that there are no significant disagreements among other international actors on AI. Still, the reality is that in Paris the European Union managed to find a mutually acceptable compromise with the Global South, India turned out to be able to overcome its frictions with China, and only the US and the UK preferred to challenge the emerging global AI consensus.

It is yet to be seen how the ongoing transatlantic disputes will be resolved. It is clear, however, that the disagreements go far beyond the AI domain, reflecting much more profound political, social and even civilizational divergences between the Anglo-Saxon and the Continental European worlds. Nonetheless, the numerous AI management challenges cannot wait until both sides of the Atlantic Ocean sort out their disagreements and come to a common denominator. The AI global governance should move on—with the US and UK governmental participation, or even without it for the time being.

Alexander Ignatov:
BRICS Expansion and Prospects for BRICS Cooperation in the Digital Economy

If a comprehensive universal and legally binding AI code of conduct is not yet within our reach, we should focus on relatively modest incremental and voluntary steps that could pave the way to addressing more sensitive and controversial matters in the future. For instance, the time has come to fight for more AI transparency at all levels, including enlightening the general public, promoting expert dialogues and demanding more information on AI plans from governments and private companies alike.

At the same time, in order to counter unavoidable algorithmic biases of diverse AI platforms one should find a way to fully integrate leading models and databases through user-friendly interfaces; for end users, GPT-4 and DeepSeek should become not alternatives, but rather mutually complementary tools. Finally, robust public oversight of AI development remains an essential, if not sufficient, safeguard against any apocalyptic “rise of the machines” scenario.

“We must all hang together, or, most assuredly, we shall all hang separately,” wisely noted Benjamin Franklin at the very dawn of the American Revolution. This logic fully applies to the existential AI challenge to the humanity that we now see at our doorsteps.

First published in the Global Times.