Au bout de 40 zettaoctets, Noé ouvrit la fenêtre | Anna Nesvijevskaia | TEDxGeneva

Quick Overview

Anna Nesvijevskaia argues that the world is drowning in data, citing that 40 zettabytes of data were generated by 2010, and this volume, combined with the inherent biases and complexity of data science, requires a shift towards AI systems that mirror human learning and creativity rather than just massive data processing.

Key Points: By 2010, 40 zettabytes of data were generated, a volume that overwhelmed traditional methods of data collection and analysis. The speaker, formerly an analyst, decided to pursue a PhD in Big Data after realizing the current data landscape was insufficient for understanding complex phenomena. Governments label scientific publications and military/financial data as strategic assets, leading to data hoarding and hindering open scientific collaboration. The speaker references the story of Noah, who waited 40 days and nights for the flood to subside before opening the window, suggesting a need for patience and reflection amidst data deluge. Modern AI, particularly generative AI, is capable of reproducing human-like thought processes and creativity, which is necessary to navigate the 'ocean of data' containing hidden patterns and insights. The speaker's father, a physicist, asked her critical questions about the nature of data, suggesting that data science requires more than just technical skill; it demands essential human elements like ethics and narrative. The talk concludes by contrasting the immense, often messy, data archives with the need for simpler, more human-understandable insights, suggesting a future where AI helps us navigate chaos.

Context: Anna Nesvijevskaia delivers a TEDxGeneva talk under the theme "CHAOS" (November 28, 2025), focusing on the overwhelming scale of data generation (zettabytes) and the limitations of current data science approaches. She contrasts the massive scale of historical data collection by civilizations like the Egyptians and Mesopotamians with the modern computational explosion, arguing that sheer volume is not the solution; rather, the focus must shift to creating intelligent systems capable of human-like intuition and creativity to extract meaningful insights.