Anurag Shekhar, Calvin Mabaso
Five themes were developed. Commenters described AI companionship as a response to harmful or disappointing human relationships; debated whether AI intimacy could be considered real; linked loneliness to weakened communities and wider social conditions; viewed constant validation as both supportive and limiting; and expressed concern that platforms profit from users' intimate disclosures. The discourse combined empathy, scepticism, concern, and moral judgement.
INTRODUCTION: Loneliness is a major social problem, and AI companionship has become a contested response. Most research has focused on individual users, while less is known about how AI intimacy is debated in public-facing digital spaces.
METHODS: This study examined 8,757 YouTube comments posted in response to a mainstream news video on AI companionship. Sentiment analysis, emotion analysis, and topic modelling were applied to the full corpus. Reflexive thematic analysis was conducted on 527 high-engagement comments.
RESULTS: Five themes were developed. Commenters described AI companionship as a response to harmful or disappointing human relationships; debated whether AI intimacy could be considered real; linked loneliness to weakened communities and wider social conditions; viewed constant validation as both supportive and limiting; and expressed concern that platforms profit from users' intimate disclosures. The discourse combined empathy, scepticism, concern, and moral judgement.
DISCUSSION: Guided primarily by Honneth's theory of recognition and Zuboff's analysis of surveillance capitalism, the study interprets AI companionship as a contested form of intimacy that may meet genuine emotional needs while remaining embedded in commercial platform structures. The findings reflect responses to one video and should not be treated as representative of public opinion. They nevertheless show how public-facing digital discourse can reveal changing social understandings of intimacy, loneliness, recognition, and platform power.