Intelligence Beyond the Algorithms: Network Shape Intelligence and the Historicity of AI

This article challenges the prevailing notion that artificial intelligence (AI) can be understood within a single, all-encompassing paradigm. While acknowledging the coherence and influence of established historical frameworks in AI, it argues that such approaches risk flattening the cultural, political, and epistemological complexity of the field. These frameworks often culminate in a historiography of AI presented as a chronicle of linear progress. Drawing on an interview with neural network scientist Carlo Vittorio Cannistraci, this article proposes an alternative historiographical model—one that understands intelligence not as a universal algorithmic potential, but as a mirror of human societies, shaped by linguistic, cultural, and ethical difference. This conception emerges from a philosophical-historical reading of Cannistraci’s notion of “network shape intelligence.” By framing AI as a reflection of historical traditions and sociotechnical imaginaries, the article advocates for a historiography “from below,” integrating insights from philosophy and history of science. The interview’s exploration of Prometheus, toolmaking, training, and international contrasts—particularly between the US, Europe, and China—underscores the need for AI research to engage with questions of freedom, responsibility, and systemic justice. In contrast to dominant trends, the article calls for a pluralistic, historically grounded, and socially engaged understanding of intelligence—whether natural or artificial. The interview thus lays the groundwork for a new historiographical approach: one that resists linear genealogies and instead opens space for rethinking intelligence through the lenses of training, education, discipline, and cultural difference. In doing so, the article argues that the historicity of AI constitutes a shared field of inquiry for both humanists and scientists.