The world race for artificial intelligence goes in parallel with a global competition to capture big data fluxes. The rationale is simple. In the new artificial intelligence paradigm, algorithms can be transparently published and shared as common goods. The real value lies in the data to train them. In this gold rush for training data, the largest companies in the US and in Asia enjoy a dominant position as they mediate and therefore capture a massive daily flow of information.
But Europe has an asset that may change the rules of the game: kilometres of archive and library shelves, immense collections from museums and other cultural heritage sources and a solidly established know-how for interpreting and analysing them. These records, being mapped by Europeana, are not just important for preserving heritage of the past, they document billions of pieces of information about our world today.
Europe is about to become the leader in the extraction and modelling of this ‘Big Data of the Past’, transforming it into a source of new knowledge. This will impact not only education, creative industries, policymaking and environmental modelling but will also permit the emergence of a new kind of Artificial Intelligence with wider temporal horizons. This future AI, not focused on the pulsations of the present, but trained on millennia of multilingual historical records could lead to a powerful new understanding of our world and its long-term patterns. It will act as a simulation engine for possible futures, thus offering a unique advantage to research institutions, decision-makers and businesses.
The AI race is not lost. It has just begun.