Giuliano da Empoli: 'The engineers of chaos multiply anger using algorithms' (www.lemonde.fr)

Giuliano da Empoli, a political thinker and former tech optimist turned critic, has published L'Heure des prédateurs (to be released in English next October as The Hour of the Predator), arguing that a new breed of autocrats has allied with tech magnates to weaponize digital platforms. Drawing on his experience in politics and his earlier faith in “start‑up” disruption, da Empoli claims “engineers of chaos multiply anger using algorithms”: engagement‑driven systems, big‑data microtargeting and behavioral profiling have been repurposed to amplify grievance, polarize publics and accelerate populist movements from Trump to Bolsonaro. For the AI/ML community this is a sharp reminder that design choices and optimization targets have socio‑political consequences. Da Empoli’s critique spotlights technical mechanisms—recommendation algorithms that optimize for attention, feedback loops that reward outrage, and microtargeting enabled by granular datasets—that can produce large‑scale manipulation. The implication for researchers and engineers is concrete: move beyond raw engagement metrics, invest in evaluation frameworks that measure societal harms, apply privacy and provenance tools, and build auditability, explainability and robustness into systems. His book reframes algorithmic harm as a political technology problem and calls for technical, regulatory and institutional responses from the AI field to prevent platforms from becoming accelerants of authoritarian influence.