Review of literature on AI influence on interpersonal communication and social relations
Description
These tables present the results of a two stage literature review for the paper entitled "CONNECTED OR DISCONNECTED WITH ARTIFICIAL INTELLIGENCE: THE USES AND ABUSES OF AI IN THE NEGOCIATION AND CONSTRUCTION OF SOCIAL RELATIONS" written by Alexandra Gheondea-Eladi.
Files
Steps to reproduce
We have performed a literature review in two stages. First, an unstructured, purposeful literature review aimed at identifying a high variation of possible impacts of Artificial Intelligence Use on interpersonal communication and social relations. By searching academic databases through an aggregate library engine as well as Google scholar, we selected 36 papers based on title and relevance to the influence of AI on human communication and relations. From these, after careful analysis, 13 contained relevant information. In the corresponding data of this analysis, we classified each of them according to whether they mention potential positive uses, abuses, connecting and disconnecting capabilities of AI use, as well as those fostering solidarity or the lack of it. The second stage comprised of a structured search in Web of Science, aimed at obtaining a reproducible analysis of the relevant literature. We used the short list of relevant keywords derived from the previous unstructured review: AI (Topic) or LLM (All Fields) and social relations (All Fields) and communication (All Fields) and conflict resolution (All Fields) and interpersonal (All Fields) and coaching (All Fields) which has been restricted to articles and to the years 2023-2027. The search lead to 246,739 results. From these, we restricted the search to those which included “social relations” and yielded 425 results. After reviewing the title and summaries, we selected 32 of these for closer review. NotebookLM has been used to extract from the 32 papers examples of positive and negative implications for social relations and communication. All entries obtained in this way have been checked by the author for accuracy of reporting, over generalization, missing information or misrepresentation of conclusions. Thirteen out of 128 entries needed corrections or additions for such errors. After this analysis 30 papers remained in the analysis pool of relevant content. The corrected table was then coded according to the six dimensions, as before.