Computer parsed history by reading news stories over 150 years

Until now, people study history, relying mainly on reading the surviving historical documents. However, even in this seemingly conservative area of ​​knowledge, digital technologies are rapidly intruding.

A team of researchers from the University of Bristol, using specially "trained" artificial intelligence, "read" 35 million articles from 100 local British newspapers, numbering 30 billion words and covering a period of 150 years. Scientists paid special attention to the analysis of major events, in particular, wars and epidemics.

Perhaps the most interesting thing that the researchers saw was the ups and downs in the field of scientific and technological progress in the period from 1800 to 1950. So, for example, the decline in the popularity of steam and the rise of electricity at the end of the 19th century with the intersection of trajectories in 1898 can be clearly traced. A similar trend is observed with the advent of the railway, which gradually began to replace horse traction.

It turned out that politicians and writers at that time had more chances to take the leading roles, in contrast to scientists and mathematicians, who, as a rule, became famous much later. Men appear in the news significantly more often than women, but after 1900 the gap began to close rapidly.

The analysis of large amounts of historical information using artificial intelligence creates a situation where "living" historians may eventually be left without work.