Cognitive Extraction and the Fight for Digital Sovereignty

• Generative AI does not create new knowledge but statistically recombines existing human-generated data.
• Heavy public investment in massive AI models is fueling a system of cognitive extraction that benefits a few private corporations at the expense of the public.
• Securing democratic sovereignty requires redistributing benefits to data creators and protecting digital infrastructure as public goods.

The Reality of Cognitive Extraction

• AI learns from human labor—including Wikipedia editors, open-source developers, and artists—without providing compensation for the value extracted.
• Much like the historical appropriation of popular art by private markets, AI models are turning collective human knowledge into private commodities.
• This process represents the largest project of cognitive extraction in history, threatening to concentrate power in the hands of a few tech firms.

Policy Path to Redistribution

• Public funding should pivot from subsidizing bigger models to taxing AI providers and supporting funds for knowledge-producing communities.
• Software developed with public funds must remain open source and auditable to prevent state dependence on private, proprietary infrastructure.
• Digital infrastructure must be treated as a common good to avoid the silent privatization of knowledge.

Technical Debt and Loss of Sovereignty

• The so-called "wow effect" of AI demos obscures the long-term reality of maintenance and the accumulation of massive technical debt.
• As developers rely on automated tools they cannot audit, they lose the ability to maintain their own systems, creating total dependency on model providers.
• The battle over AI is fundamentally democratic, forcing a choice between the concentration of power and the achievement of cognitive justice.