By Marcel Timmer, country manager at Red Hat Netherlands
Digital sovereignty is one of the most important tech discussions of the moment. It involves not only who controls data but also the autonomy of nations, organisations, and individuals in the digital space. DIgital sovereignty means full control over your own digital footprint; it means avoiding lock-in; but it also means being able to fully understand all interactions between internal and external systems. So that is the crux of the story: control and protect data to remain agile and resilient. But why is this such a challenge – and what is the solution?
The Importance of Digital Sovereignty
The adage “knowledge is power” also applies in a digital world. And while knowledge was previously primarily stored on paper or other physical media, its digital counterpart, ‘data’, has become somewhat intangible for many people; it floats around in the ‘cloud’. Essentially, it is of course physically stored on servers, but thanks to the internet, digital data is more accessible to more people than ever before.
This accessibility brings many advantages, but it also makes data a target for criminals and hostile governments, among others. To prevent data from being stolen or altered, control is needed – sovereignty over the data. A lack of control exposes people and organisations to numerous risks, from identity theft and financial fraud to loss of lives.
The issue of digital sovereignty plays out at different levels. For instance, the control that citizens have over their own data is at the heart of discussions around tracking cookies and the activities of parties that have collected personal data for facial recognition without permission. For organisations, it often revolves around securing business-sensitive information and retaining the freedom and flexibility to be independent of specific commercial suppliers. For governments, digital sovereignty is not only protection against espionage and digital warfare but also, for example, against interference by other governments in domestic policy.
The Path to Self-Determination Is Open
In all these perspectives on digital sovereignty, ‘self-determination’ is the common denominator, as NLDigital notes in its position paper. And open technology is essential to secure that self-determination. Enterprise open source, in particular, can offer a level of transparency and flexibility that can be lacking in commercially licensed software. Enterprise open source software is supported by a vendor that stabilises and quality-assures the software, certifies it works with an ecosystem of hardware and software and secures it. This leads to more visibility over software supply chains and a greater capability for quick adjustments to market changes and regulations.
The underlying infrastructure must also be flexible and open enough – hence why increasingly more organisations are moving to a hybrid environment. A hybrid multi-cloud that encompasses both public and private cloud environments reduces reliance on a single cloud service provider.
Sovereign AI?
A relatively new aspect of the sovereignty discussion revolves around AI. The majority of widely used AI solutions are developed outside the EU; the data they are trained on is not always transparent, and the norms and values these AIs are endowed with do not automatically align with the organisation using them.
Therefore, it is not very surprising that European countries, including the Netherlands, are looking at developing their own AI. This is driven in the private sector by companies like the German Aleph Alpha, which recently formed a partnership with the Finnish company Silo to develop European AI solutions.
Here in the Netherlands, the government is encouraging the development of its own large language model through the initiative GPT-NL. 13.5 million euros have been allocated for this. The question is whether that is sufficient. Currently, AI development is reserved for very wealthy organisations. It can cost up to 1 billion dollars to build and train a new generative AI model and even more to use it on a large scale. Access to the costly chips that drive these AI systems is controlled by a few companies from Silicon Valley. Also consider that the required knowledge is very scarce. Much technical knowledge is widely spread, but other skills, especially in scaling AI system training, are possessed by only a few thousand people worldwide.
Perhaps the way forward does not require a national focus but international collaboration. A ‘federation’ of sovereign AI initiatives could pool their individual, medium-sized computing capacity so they can collectively train a model with the benefits of large-scale computing power. Another option is ‘Swarm AI’, where a large number of small, task-specific models are deployed instead of large commercial LLMs. Small models are more dynamic, cheaper, transparent and traceable. And they reduce dependence on specific suppliers. All this is in line with European regulations such as the AVG and the AI Act, which emphasise transparency and reliability.
Conclusion
Digital sovereignty is not a luxury, but a necessity in today’s digital landscape. It is about ensuring control, security, and autonomy in a world that increasingly relies on data and technology. Open technology, hybrid cloud environments, and, in recent years, initiatives for sovereign AI are crucial steps towards a future where individuals, organisations, and nations have self-determination over their digital existence. It’s a challenge that requires collaboration and innovation, with high stakes: a digital landscape that is not only advanced and connected but also safe, transparent, and under control.
Read: Red Hat announces latest updates to global partner engagement experience