Affective Impacts of Sustained and Emotionally Intense Human–AI Interaction: A Scoping Review
Xenia Armstrong · 2026
This scoping review maps empirical research on the affective outcomes of sustained or emotionally intense interactions between humans and conversational AI (CAI) systems. Following the PRISMA-ScR guidelines and the JBI methodology for scoping reviews, the review synthesizes findings across four research questions: (1) which affective impacts are reported, (2) how these impacts are theoretically explained, (3) which methodological approaches are used, and (4) what conceptual and empirical research gaps remain. Scope definition: The review's scope is defined using the Population–Concept–Context (PCC) framework: human users interacting with conversational or social AI systems (Population); affective/emotional outcomes of such interaction (Concept); sustained or emotionally intense human–AI interaction in non-technical, user-facing contexts (Context). Eligibility criteria: Included studies are primary empirical studies (quantitative, qualitative, or mixed methods) examining affective or emotional outcomes in users interacting with conversational or social AI systems, published in English as peer-reviewed journal articles or conference papers between January 2019 and December 2026. Studies focusing solely on technical performance, task-based use, or hypothetical/vignette-based scenarios were excluded. Database search strategy: A systematic search was conducted across six databases (ACM Digital Library, APA PsycNet, IEEE Xplore, PubMed, ScienceDirect, Scopus). Screening procedure: Records were imported into Zotero and screened independently by two reviewers at the title and abstract level, followed by full-text screening against the predefined eligibility criteria. Discrepancies were resolved by a third, experienced reviewer. Data charting framework: Data were extracted using a structured charting template covering bibliographic details, context and population, AI system characteristics, interaction characteristics, study design and methods, affective outcomes, measurement of affect, outcome directionality, key findings, risks and negative effects, and limitations.