News Media Robot: Theoretical Aspects of Intellectual System for Content Generation

摘要

The article discusses a new phenomenon of digital media sphere, i.e. news media robots, which are capable of creating content on a technological platform of intellectual systems. Interdisciplinary approaches to understanding this phenomenon, which attract theoretical concepts and cases from artificial intelligence, knowledge engineering and natural language technologies are considered. It is stated, that in the core of intellectual systems lie computational models of natural human intelligence and knowledge. The phenomenon of imitation as a characteristic of social and technical systems general ability to self-study relying on a sample is viewed as a key mechanism media robots adaptation to social environment. The case study of generating news content within the context of frames theory is provided as an illustration of universal formalism for representation of knowledge and social actions. The structure of general human intelligence, provided by Cattell — Horn theory, is considered as a possible reference model for comparison of a robotic intellectual system with natural human intelligence. The similarity between a general intelligence structure and an “intellectual” structure of a media robot is noted. The article defines the technology of modelling a natural language as the most prospective means of broadening the verbal intellect of a media robot. It is stated in conclusion, that one can observe trends for convergence in the methodology of knowledge engineering and analytical journalism. In a long-term perspective serious progress in creating new generation of media robots requires reproduction of such human abilities, as learning and “sense” of social responsibility. For this reason joint efforts in interdisciplinary studies in digital media, artificial intelligence and cognitive sciences are required.