If the preceding Industrial Revolutions went hand in hand with the nation-state, the ongoing Fourth Industrial Revolution appears to bypass the state altogether. 

Defined by staggering advances in Artificial Intelligence (AI), which has replaced computing power as the “game in town”, the world is on the cusp of the “unknown unknown”. 

In this paradigm, even Moore’s Law, which postulated that “the number of transistors on a chip doubles every two years, while the cost of hardware (computers) is halved”, seems to have been struck by the gale of “creative destruction”. 

But amid the AI ‘blitzkrieg’ that is leaving past techno-economic paradigms as ‘roadkill’, there are now growing calls for “AI sovereignty”. 

A pale echo of “industrial sovereignty”, a formulation dating back to the heyday of decolonization, “AI sovereignty” is a defensive concept. Best exemplified by the decision of the U.S. tech firm Anthropic, spurred by an injunction from the U.S. Department of Commerce, to stop global access to its Fable V and Mythos 5 models, “AI sovereignty” is the new buzzword.

Does the formulation mean anything, or is it merely a rhetorical call to arms by the Third world? Can the AI juggernaut be stopped? Should it be checked? Should the state cede space to AI in an “after-nation” condition, where technological determinism leads to a tech “nirvana”? 

And if the U.S. closes its proverbial doors on AI, AI stack development, and the models that underpin it, what options will the rest of the world be left with? Should it turn to Europe?

In essence, “AI sovereignty”, an appealing term for the sensibilities of the decolonized Global South, is a hollow phrase. The reasons pertain to the very nature of AI. It is a juggernaut that cannot be reified. It is neither entirely tangible nor capable of being “captured”. 

While the initial foundations of modern technological development, the bits and bytes of computing and the internet, emanated from the Pentagon and benefited from industrial policy, AI development is, more or less, autonomous. 

In a loose sense, AI sovereignty may imply the control and delimitation of AI within the confines of a given nation-state. 

But given the nature of the technology and its development trajectory, it is difficult, if not practically impossible, to “capture” AI. 

If that is the case, it leads to an obvious corollary: in the race for AI development, stack-building and model innovation, collaboration is key. 

But collaboration, like any powerful tool, can be exploited. Under the banner of partnership and AI as a global public good, it can enable misuse, free-riding and monopolization. 

In an era where AI is already eroding the power of patents and intellectual property, what safeguards remain to prevent innovation from being copied or........

© Kashmir Observer