The rise of data colonialism marks a shift in global power structures, where data is extracted, managed, and monetized by a few dominant tech companies, reshaping democratic processes and societal norms. This concentration raises urgent questions about digital sovereignty, transparency, and the role of platforms; as Europe begins to assert itself with laws addressing deepfakes and AI governance, the debate intensifies around who controls data and how it is used. The challenge lies not only in protecting personal data but also in rethinking the social contract, fostering distributed governance, and ensuring technology evolves within ethical, human-centered frameworks.

Alberto Barachini stated that data is now a power tool that can influence how and where information is consumed. “News”, he argues, “is a fragile and impactful product, and must be protected. These companies are acting as editors without editorial responsibility”. Supporting the publishing sector, once seen as eccentric, is now becoming a model. Silvia Castagna highlighted how the ecosystem itself is fragmenting: on one side, traditional publishers; on the other, private individuals and aggregators who monetize attention without producing news. “The data is us”, she warns. Marco Trombetti traced this disruption back to language itself. Starting from translation models, AI evolved into Q&A systems trained on the top 1% of internet content. “We broke the internet”, he says, noting the shift to a model where AI absorbs and learns from interaction itself. Jannis Kallinikos sees in this a deeper change in the social contract: data is the outcome of designed engagement, where platforms shape the economic organization. In this distributed, contingent system, only distributed governance can hold.

Speakers highlighted that the first wave of AI training may be over, but the conversation is only just beginning. While the foundations were laid without much oversight, there is now a growing awareness that regulation must catch up. And yet, as AI begins to learn through interaction, the nature of data itself is shifting. It will no longer be a matter of what data we have, but of how we choose to shape the systems that learn from it, and whether we can still steer them in line with the values we wish to protect.