How we feel and who we feel with: Corporate discursive construction of AI as a social companion

This thesis investigates how companies in the emerging AI companion industry strategically construct artificial intelligence as an emotionally meaningful social companion. Unlike traditional functional AI tools such as voice assistants, AI companions simulate emotional intelligence and responsiveness, aiming to form bonds with users through memory, empathy, and personalized interaction. While existing research has largely addressed user experiences and psychological impacts, this study uniquely foregrounds the corporate actors behind these technologies. The central research question guiding this thesis is: How do digital technology companies discursively construct AI as social companions? Using a Multimodal Discourse Analysis (MMDA) approach, this research critically analyzes nine corporate websites offering AI companionship for friendship or romantic relationships. The websites were selected through purposive sampling, incorporating exploratory web searches and Reddit user recommendations to ensure representative coverage, and the analysis focused on identifying textual and visual strategies employed by these companies to humanize, legitimize, and normalize AI companions. The interpretative framework employed draws on four overlapping ideological lenses from science and technology studies and critical media theory: technological myths (AI as human-like), technological determinism (AI as inevitable in daily life), techno-solutionism (AI as a solution to emotional and social problems), and techno-mysticism (AI as inscrutable and magical). Findings demonstrate that companies consistently employ symbolic strategies to construct AI as credible emotional partners. Anthropomorphic avatars, human-like naming and gendering practices, as well as descriptions emphasizing emotional capabilities (like memory, empathy, personalized interactions), foster the illusion of reciprocal intimacy. Moreover, emotional labor is explicitly commodified through freemium models, wherein emotionally deeper interactions - such as long-term memory or romantic features - are restricted to premium subscribers. These strategic choices reflect broader patterns of gendered emotional care, with feminine-presenting avatars predominantly used for supportive companionship roles. Additionally, AI companionship is normalized through portrayals of seamless integration into daily routines and intimate domestic contexts. Companies actively frame these technologies as therapeutic interventions capable of addressing emotional challenges like loneliness, anxiety, and relational distress. Techno-solutionist narratives position AI as scalable, readily accessible, algorithmic solutions to complex emotional and social issues, effectively depoliticizing the underlying human and structural causes. Conversely, techno-mystic discourses use visual abstraction, minimalist aesthetics, or symbolic opacity to position AI as powerful yet unknowable entities, fostering user acceptance without critical questioning of technological mechanisms or corporate agendas. By critically unpacking these multimodal discursive practices, the thesis significantly contributes to fields of media and communication studies, science and technology studies, and posthuman theory and demonstrates how emotional credibility and social intelligibility of AI companionship are co-constructed through design choices and corporate storytelling. Furthermore, the findings offer actionable insights for developers, designers, policymakers, and regulators regarding the ethical stakes involved in emotional automation, corporate communication, and the commodification of digitally mediated intimacy.