define Sovereign AI.
Jensen Huang, CEO of NVIDIA, has been an advocating for the development of Sovereign AI across various nations. In an event on Feb 2024 at the World Governments Summit in Dubai, he quoted “Sovereign AI codifies your culture, your society’s intelligence, your common sense, your history — you own your own data”. In his past interviews, he emphasized that we are “currently witnessing the significant second wave of AI” which includes the acknowledgment that “every region and every country must develop their sovereign AI” due to the distinct language, cultural requirements, and unique business capabilities of individual nations.
What does AI mean for sovereignty?
Nation's sovereignty is a political concept, that is not strictly defined and sometimes confusing. It concerns territory and borders, natural and digital resources ‘that belong to us’, the people, and authority that has internal and external legitimacy.
- Internal legitimacy pertains to the state’s efficacy in serving its citizens, including the provision of functional public services and equitable justice.
- External legitimacy involves recognition by foreign nations and a nation’s autonomy in its dealings with others.
When AI deployment overlooks broader societal systems like public service delivery, judiciary, and political accountability, it jeopardizes the legitimacy of authorities and government.
AI can both strengthen and undermine these forms of legitimacy.
Therefore, what’s the solution?
A single solution often proposed to address the complexities of AI is “transparency.” However, AI comprises intricate socio-technical elements, including data, hardware/software architectures like cloud, algorithms, decision systems, and the expertise of designers, coders, and providers. These components interact and influence each other, making achieving end-to-end transparency challenging. Explaining the workings of deep learning algorithms is already difficult, and questions about the integrity and bias of training data arise. Ensuring data confidentiality and uncovering the socially constructed aspects of AI further complicate transparency efforts. Achieving transparency in AI proves easier said than done. There is a long way to go and lot to be done, however the future looks promising.
How can we ensure the protection of democratic principles and legitimate sovereignty amidst the rise of AI?
Initially, it’s crucial to acknowledge the obstacles and increase awareness regarding the potential threats posed by AI, its deployment methods, and anticipated outcomes. Emphasizing comprehensive accountability and transparency in AI is vital, though acknowledging their practical and potentially inherent constraints. Furthermore, the challenge surpasses individual unethical developers, social media firms, cybersecurity agencies, and governments.
Final Thoughts
Acknowledging the international aspect is crucial, both in recognizing challenges and seeking solutions. While attaining global consensus on AI’s relationship with sovereignty is daunting, it shouldn’t hinder collaborative efforts to address shared concerns like autonomous weapons while advancing AI’s benefits in sectors like healthcare, agriculture, and education. A framework of rules, whether internationally, regionally, or nationally established, should consider the global landscape.
Reference: Timmers, P., 2021. AI challenging sovereignty and democracy. TPQ, 20(4), pp.45–55.
