Abstract
Artificial Intelligence (AI) is reshaping how human and non-human actors connect, organise and narrate social change. Despite the growing use of AI-driven communication technologies, empirical research remains limited on how emotional framing operates when collective action is mediated by automated agents, particularly within social movements. For South African social movements, this shift is not only technical but deeply emotional, shaping how hope, anger and solidarity are communicated in struggles for justice. While social bots are widely deployed in politics and marketing, their potential to support activism and collective action through emotional framing remains underexplored. This study addresses this gap by examining how emotional framing functions as a strategic communication tool through AI-powered social bots within South African social movements. Anchored in theories of social movements, emotional framing and strategic communication, and informed by affective computing, the study adopted an interpretivist qualitative two-phase design. Phase 1 drew on 17 in-depth semi-structured interviews with grassroots activist practitioners, campaign and media strategists, and AI and social bot developers, all recognised as strategic communicators. This phase explored how emotion is intentionally designed, governed and managed in digital activism. Phase 2 involved semi-structured interviews with 10 AI-powered social bots developed for causes such as gender-based violence, youth empowerment and refugee advocacy, examining how emotion is simulated and adapted through algorithmic communication. The integrated findings demonstrate that emotion remains a central driver of mobilisation, even when mediated by machines. When thoughtfully designed and ethically governed, social bots function as affective collaborators that humanise digital spaces, sustain engagement and extend activist reach. The study proposes a context-sensitive emotional framing framework structured around four dimensions: emotional design, ethical governance, algorithmic adaptability and infrastructural inclusivity. Theoretically, the study extends emotional framing and strategic communication theory by repositioning strategic communication as a posthuman, transdisciplinary practice in which human and non-human actors co-create meaning. Practically, it offers policy-relevant and managerial guidance for activists, developers and policymakers on designing emotionally resonant and socially responsive AI communication in unequal Global South contexts. In doing so, the study contributes original empirical and conceptual insights to scholarship on emotional AI and digital activism within South Africa and beyond.