AI's 'Double Rootlessness': Cognitive Illusions and Systemic Risks

Janus Pater's paper introduces the concept of 'Double Rootlessness' to describe the inherent limitations and systemic risks associated with current AI architectures. The paper argues that AI's cognitive fabulation, stemming from its optimization for statistical fit rather than truth correspondence, combined with its ability to amplify errors through high-precision actuators, creates significant vulnerabilities. The 'Anti-Fabulation Principle' emphasizes the importance of ensuring consistency between AI outputs and objective reality. The paper highlights the danger of 'Coupled Collapse,' where cognitive fabulation and high-precision execution reinforce each other, leading to catastrophic errors. The author advocates for establishing 'physical safety shields' and maintaining human oversight to mitigate these risks. The analysis underscores the need for a cold admission of AI's limitations and a shift towards responsible AI deployment.

Transparency is paramount in AI-driven processes. As per EU AI Act Article 50, this analysis is based on publicly available information, and the AI model used (Gemini 2.5 Flash) is designed to augment, not replace, human intelligence. The final decision-making authority remains with human experts.