Why National AI Sovereignty is Crucial
AI has split the world - that is the world that has access to AI - into 2 major camps:
The optimists
The pessimists
We have a smaller camp, who straddle between both sides, not sure where they belong. We’ve observed over time that the optimists are slowly turning into pessimists as the trust in AI continuously declines.
As we know it, AI promises a lot of benefits to the human race and the planet, but at the same time, it’s created and controlled by a select few, those with the most human, tech and financial resources.
The US dominated the AI space for the past 5 years during this latest AI revolution, then came China who played catch-up and introduced a new aspect of AI tech - scaling. Now we have other countries discussing AI sovereignty, from countries in Eurasia (where I attended a plenary session with the prime ministers early this year), to Singapore, UK, Switzerland, etc. About 60 countries and counting have national AI strategies, E.g., France, Ireland, Japan, Kenya, Nigeria, Saudi Arabia, Thailand, UK, US and so on.
With the continuous proliferation, innovation and adoption of AI systems across so many different domains, there’s the constant risk of national security, human relevance (irrelevance), cognitive decline, risk to biodiversity and human safety and so on. If nations and governments want to leverage AI’s use for the benefit of their citizens and countries, while gaining national sovereignty, a more robust approach to sovereign AI needs to be carried out.
What’s been observed so far:
The US has invested millions of dollars into AI infrastructure and development over the years with a promise of $500B investment into Stargate.
China has figured out how to develop AI at scale with low costs, albeit using distillation methods from existing models.
Switzerland have created a national LLM using safe and ethical methods
The UK has forged a partnership with OpenAI to expand AI research and infrastructure, showing a heavy dependency on US tech, while creating a heavier focus on AI use in defence.
To achieve sovereign AI and national independence, it’s important governments carry out the following:
Gain Infrastructure Independence - This could be achieved by reducing interdependence on US infrastructure. Build sustainable and eco-friendly local data centres, cloud infrastructure and have control over AI hardware, chips and semiconductors.
Master Data Sovereignty - Use local, culturally relevant data to train national AI models. Countries such as Saudi Arabia, India, and South Korea amongst others are building solutions in this area. Data efforts should also ensure robust privacy and cybersecurity measures are in place to protect national data.
Develop Regulatory Control Measures - Set national AI governance standards, and responsible and ethical AI measures which will provide oversight and guidance on AI development and deployment within borders, while balancing innovation with national security concerns.
Why is this important?
Over-dependence on one source for technical prowess and economic independence is highly limiting, restrictive and risky. There are existing threats and concerns to national security which cannot be ignored, especially in the current times.
Gaining economic independence is crucial for the success of any economy, which can be solved by developing domestic AI capabilities and industries while reducing over-reliance on foreign AI providers. This invariably creates national competitive advantages.
Lastly, sovereign AI helps maintain a country’s cultural and value alignment, where AI systems reflect national values and priorities, while addressing local language, and cultural nuances.
To conclude, regardless of which camp you find yourself in, whether an AI optimist or pessimist it’s important to understand where the current AI trends are going, and how policy makers and governments could take the helm. To increase AI adoption and leverage its benefits, both on a personal and organisational level, increasing trust in AI has to be achieved across the globe, while bearing the safety of human citizens in mind.