Session 51. What happens when massive computing power brings together an ever-growing cross-section of the world’s information in realtime, from news media to social media, books to academic literature, the world’s libraries to the web itself, machine translates all of that material as it arrives, and applies a vast array of algorithms to identify the events and emotions, actors and narratives and their myriad connections that define the planet to create a living silicon replica of global society? The GDELT Project (http://gdeltproject.org/), supported by Alphabet’s Jigsaw (formerly Google Ideas), is one of the largest open data initiatives in the world focusing on cataloging and modeling global human society, offering a first glimpse at what this emerging “big data” understanding of society looks like. Historically GDELT focused on the textual world, mass machine translating everything it monitored globally in 65 languages, using sentiment mining to assess 4,500 different emotions, geocoding mentions of location down to the level of a hilltop, identifying millions of topics and hundreds of event classes and making it all available as an open dataset available through BigQuery and GCS. Yet, imagery and video are increasingly the way global events and narratives are communicated and so we needed a way to go past the text to catalog world events through the eyes of all of that visual material – enter Cloud Vision. Over the past 17 months we have cataloged more than a quarter billion news images from every corner of the globe, using Cloud Vision to literally watch the world go by each day. From mapping what “land vehicles” look like in every corner of the world to realtime assessment of air and land pollution from the backgrounds of those images to live alerts of flash flooding, earthquakes and other natural disasters with ground truthed estimates of damage severity, to mapping how “happy” or “sad” the news imagery of the world is to differing violence norms to peering into the world’s visual narratives (what are the images picked by each country’s news outlets to discuss Donald Trump?) Together with early experiments using Cloud Speech and Cloud Natural Language, this talk will dive into how we are using Google Cloud and especially its AI APIs to catalog our world in realtime