Given that artificial intelligence (AI) technology has now spread widely in the field of new media, personalised recommendation algorithms have reshaped the pattern of information distribution; at the same time, they have also given rise to or worsened the problem of “information cocoons”. AI-driven information cocoons are chosen as the research objects, and in combination with theoretical analysis and case studies, their basic concepts and attributes have been defined, formation mechanisms under the three dimensions of technical logic, commercial logic and user psychology have been examined, and their adverse effects on individual cognition and social ecology have been investigated. Based on the above, three types of breaking strategies at the levels of government regulation, industry self-discipline and individual empowerment have been proposed to provide a theoretical basis and practical path for building an open and diverse new media information ecosystem.
