Until a few months ago, the debate on artificial intelligence centred on language models: who had the most powerful, the fastest, the most accurate, and so on. Today, however, the most significant news stories are about the cloud, data centres, platforms and integration with operating systems.
Competition is no longer solely about the quality and reliability of the models, but about the infrastructure through which that intelligence will enter people’s lives.
And that is where the future will be decided: who will influence the way in which information reaches us, how it is interpreted, understood and used? How, then, will we shape our knowledge?
For years, the debate on European digital sovereignty has centred on a simple yet fundamental question: who controls our data?
It is this perspective that has guided a significant part of European digital policy. Europe, often accused of being over-regulated, has introduced the GDPR, which has become a global benchmark for the protection of personal data, the Digital Markets Act and, more recently, the AI Act; the latter two have sought to introduce rules in a market dominated by global technology platforms.
Generative artificial intelligence is shifting the focus even further: the issue is no longer simply about where data is stored or who can access it, but the question becomes more complex and centres on who controls the tools through which that data is transformed into a narrative.
It is here that a new dimension of digital sovereignty emerges: cognitive sovereignty
This concept requires us to look beyond the mere availability of information and demands that we retain the ability to understand, interpret and guide the processes through which that information first becomes answers to our questions, then representations of reality, and finally decisions.
From information to knowledge: the paradigm shift in artificial intelligence
Digital communication over the last twenty years has had one constant: search engines. Google and the other major players have organised the vast amount of information available on the web, making it accessible to users; for businesses, this has meant developing increasingly sophisticated strategies to gain visibility: content optimisation, SEO, search ranking and online reputation management.
The fundamental question in digital communication has long been:
“How can I be found?”
Of course, there was a downside. A company’s reputation could be jeopardised in a matter of hours; when a negative incident occurred, the top results on search engines would be flooded with content sourced from newspapers and press agencies, and simply managing to get a company’s statement to appear amongst the top results was already a success.
Today, however, generative artificial intelligence is introducing a different model: users no longer ask ‘Where can I find this information?’, but increasingly ask: ‘Give me an answer’.
And whilst the answer isn’t always accurate, the process is certainly very straightforward. The user experience is so effective that we risk accepting it without questioning the quality of the process that led to that answer. Often, even the links provided as sources for the answer seem redundant, and the issue of source accuracy – on which a digital oracle like Wikipedia has built its philosophy and success – risks becoming irrelevant in the face of our own laziness.
The focus therefore shifts: from research to synthesis, from access to interpretation. It is precisely this shift that is the most significant.
An artificial intelligence system does more than simply retrieve content: it selects it, connects it, assigns it a weighting and constructs a narrative. The knowledge intermediary is evolving from a source recommender into a new ‘content creator’ that tells us about the world.
But what criteria determine what an artificial intelligence considers relevant and reliable for us?
The competition is shifting from algorithms to ecosystems
Competitive advantage stems from the integration of various elements, such as computing infrastructure, the cloud, data, artificial intelligence models and applications, right through to the platforms through which users access services.
The collaboration between giants such as Microsoft and OpenAI, the integration of AI into Google, Apple and Amazon products, and Nvidia’s investments in computing infrastructure show that big tech companies have moved beyond the technological challenge and are now competing on their ability to build comprehensive ecosystems.
Those who control these ecosystems do not simply control a tool, but an ever-increasing part of the processes through which information and knowledge are produced, distributed and used.
From digital sovereignty to cognitive sovereignty
The issue is not just about who possesses the technological resources needed to process information, but about who sets the criteria by which that information is interpreted, organised and transformed into knowledge.
This is the phase we are currently going through.
In a world increasingly shaped by intelligent systems, the risk is no longer simply that of losing control over data, but of gradually losing the ability to understand how the responses that guide perceptions, behaviours and decisions are constructed.
The challenge has shifted from preventing knowledge from circulating through global ecosystems to ensuring that these ecosystems are transparent, pluralistic and capable of giving due consideration to different perspectives.
Cognitive sovereignty means retaining the ability to participate in the construction of the systems through which we interpret reality, because those who influence the processes of selection, synthesis and distribution of knowledge exercise a new form of power: not necessarily the power to possess information, but certainly the power to determine its meaning.
Communication enters the era of artificial intermediaries
This transformation will also have far-reaching consequences for the world of communication.
The question now is: ‘How can I be recognised as a reliable source by a system that interprets information on the user’s behalf?’ It is a significant change.
Digital reputation will no longer depend solely on online presence, engagement or search engine rankings; it will also depend on the ability to be correctly understood by the algorithms that will increasingly mediate the relationship between organisations and people.
This raises new questions about the sustainability of digital communication.
When we talk about digital sustainability, we are not just thinking about the environmental impact of technology – such as the energy consumption of data centres, the efficiency of infrastructure and the size of digital content – but also its social and economic impact. Whilst all digital content consumes resources, it also has an effect on the information ecosystem: it influences what information becomes visible, what knowledge is valued, and which narratives gain authority.
How will our society change if it is shaped by algorithms that are black boxes?
It is a sort of ‘cognitive footprint’ of digital communication: whilst environmental sustainability measures the impact of technology on the physical ecosystem, cognitive sustainability concerns its impact on the knowledge ecosystem.
Cognitive sovereignty as Europe’s new frontier
The next phase of digital transformation will be shaped by our ability to develop more powerful technologies and to build ecosystems in which innovation, competitiveness and democratic values can coexist.
This is why cognitive sovereignty should be a fully-fledged part of the debate on digital sustainability. It is not just about technology; it concerns the potential for artificial intelligence to enhance people’s critical thinking skills, promote pluralism of information, make decision-making processes more transparent, value cultural and linguistic diversity, and prevent knowledge from becoming a resource controlled by a handful of global players.
The major challenge facing Europe, arising from the advent of artificial intelligence, is to help shape the model of society that will emerge from this transformation.
