Artificial intelligence, particularly in the form of large language models, has been at the forefront of news for several years. The majority of historians, when they express an opinion on the matter, seem mainly concerned about whether their students will be able to have essays composed by an electronic servant, and if so, to whom the obtained grade should be attributed. It is true that sometimes one looks at the finger pointing towards the star and not at the moon itself. Indeed, it seems that the impact is not so much of the invention, but of the accessibility of these devices that could be significant. The acquisition of historical data, its manipulation, as well as its interpretation are expected to take place within a workshop, with deeply renewed tools. The collapse of the cost of data acquisition and manipulation opens up multiple possibilities, among which are the potential for data-driven research, the manipulation of corpora of unprecedented size, and, paradoxically, a distancing from the text, the basic element of the historian's diet since the origins of the discipline. We will explore these potentialities through a few examples.
Philippe Rygiel is a Senior professor at Ecole Normale Supérieure de Lyon.
14.00 - 15.00
C²DH Open Space (4th floor, Maison des Sciences humaines) and online.
