Futurity as Infrastructure: A Techno-Philosophical Interpretation of the AI Lifecycle

AI systems create self-reinforcing cycles that current regulations don't adequately address: The paper identifies "futurity"—a dynamic where more data improves AI performance, which attracts more users, which generates more data—as a fundamental characteristic of AI that existing regulatory frameworks like the EU AI Act fail to account for. This recursive loop concentrates power in the hands of companies with the largest data collection infrastructures.

Data infrastructure, not just data itself, is the real source of competitive advantage and power: The paper highlights "feature stores" and other data infrastructure systems as critical but overlooked components of AI systems. These technical systems enable continuous feedback and adaptation, meaning that controlling the infrastructure—not just owning data—is what creates lasting competitive advantages and power asymmetries in the AI industry.

Current responsible AI frameworks miss the "becoming" of AI systems: The paper argues that policymakers focus on static snapshots of AI systems (training data, model outputs) rather than understanding AI as an ongoing, evolving process. This temporal blindness means regulations cannot effectively govern how AI systems continuously change and adapt after deployment through feedback loops and retraining.

Tech oligopolies are structurally entrenched through AI infrastructure: The paper identifies an escalating power asymmetry where large technology companies' control over data capture, training infrastructure, and deployment systems creates barriers that smaller competitors cannot overcome, effectively locking in market dominance in ways that traditional competition policy struggles to address.