Journalism Featuring Artificial Intelligence

Abstract

The article deals with mass media techniques based on the algorithms of artificial intelligence, principles of machine learning and deep learning based on neural networks, and recommenders. The authors provide an analytical review of experiences of applying these techniques in various spheres, namely, data processing and analysis, automated production of reports on current events and facts, interactive communication with audience, tracking newsworthy events, fact-checking, visual discovery, video content production, and others. The mass media techniques are demonstrated by the list of examples of Reuters News Tracer, Wordsmith, Heliograf, Perspective (an interface of applied programming), Newswhip, Quackbot, Guardian Chatbot, Wibbitz, Factmata, et al. Factmata is given special attention to as a complex approach to algorithmization of the Media that includes such methods as contextualization of statements, arguments and stories, and keeping a blacklist of domains which the algorithms mark as hateful, hyperpartisan, toxic, or fake news. The authors note that machine learning of algorithms for generating and analyzing texts is becoming easily accessible. Moreover, the new generation of algorithms based on artificial intelligence is able to identify text sentiment. The analysis of the impact of the media environment, including such factors as echo chamber and filter bubble, on information users shows that information can now be compared to a drug which is almost legal and easily available for use by any social group, and its users, due to targeting and personalization, are transforming into its «ideal consumers».