In the Times, the robot wrote the earthquake news

Just three minutes after the California earthquake on March 17, the LA Times published a note on the subject. The message looks quite ordinary: the exact time, the magnitude of the earthquake and its radius are indicated. The only sign of something unusual is the postscript at the end: "The post was created by an algorithm written by the author." In other words, the article was written by a robot.

The QuakeBot algorithm for processing earthquake information is far from the only program that generates news texts. In the same LA Times, there is the Mapping LA project , where you can create a text report with a comparison of different areas. The Homicide Report program automates the publication of crime killings reports.

In some areas, journalist robots have already reached perfection. For example, according to a study published a month ago in Journalism Practice, most readers are not able to reliably recognize a sports report written by a robot, writes New Scientist. Readers call such texts “credible” and “informative,” although “slightly boring.”

Robots do an excellent job with short materials on specialized topics when sources of information in the specified format are known in advance. In the fast processing of large amounts of information, they have a clear advantage over humans. In the case of QuakeBot, the text generation process was reduced to simply extracting the numbers from the US Geological Survey mailing list and filling out a pre-prepared text template with the subsequent automatic upload of the article to the newspaper’s website. The robot sent letters to the editors about the publication of the note so that they could check the text.

Programs have another advantage: they can generate millions of news notes, attracting search traffic. For example, algorithms from Automated Insightsover the past year, 300 million articles were published. An ordinary journalist seeks to write one article that as many people as possible will read it. Robots have the exact opposite task. For example, it can be text reports about the games of a fantasy sports simulator - reports that are interesting only for two owners of virtual teams. Two readers is already great.

According to experts, in the future, journalistic robots may find application in new niche areas, such as lifelogging and generation of reports from physical activity sensors, which are now very popular in the West as part of the universal passion for quantification ( collecting detailed statistics about your life in passive mode) .