Artificial Intelligence in Historical Research: Two Waves of 40 Years Evolution

Borodkin, Leonid Iosifovich

Corresponding Member of the Russian Academy of Sciences, Doctor of Historical Sciences, Professor, Head of the Department of Historical Informatics of Lomonosov Moscow State University, Head of the Center for Economic History of Lomonosov Moscow State University, Moscow, Russia
lborodkin@mail.ru

Abstract
The development of an interdisciplinary field related to the use of computers in historical research has gone through several stages over 60 years, radically changing the expectations of possibilities of using mathematical methods and digital technologies by historians. In many ways, the change of these stages was determined by the progress of computer technology. With the beginning of the 21st century, the time has come for a noticeable update of the tools suitable for processing and analyzing data of historical sources (when the data became “large”).
The paper examines the experience of using artificial intelligence (AI) technologies by historians, discusses their capabilities and limitations, taking into account the specifics of historical research. If at an early stage approaches that did not require high-performance computers (for example, cognitive methods for analyzing historical texts and expert systems) dominated, then in the last decade much more attention has been paid to artificial neural networks, the training of which requires more powerful computer resources. Examples of projects are given where the use of AI technologies has allowed historians to obtain a significant increase in knowledge.

Keywords: artificial intelligence, data science, historical research, cognitive methods, history-oriented expert systems, artificial neural networks, machine learning