Between data and dignity: Social workers’reflections on artificial intelligence (AI) use in crisis intervention

INTRODUCTION: This study aimed to explore how social workers experience and make sense of AI use in crisis intervention, focusing on what AI misses and how practitioners respond. METHODS: Using a descriptive-phenomenological design, 12 Zimbabwean social workers with international experience in technologically advanced settings in the UK and the US) were purposively sampled. Data were collected through semi-structured interviews and analysed using Giorgi’s phenomenological method, supported by MAXQDA software. FINDINGS: Findings revealed five key themes: (1) AI’s inability to capture relational knowing in emotionally charged encounters; (2) discomfort with the opaque logic behind algorithmically generated outputs; (3) tensions between data efficiency and human dignity; (4) reassertion of professional judgment through ethical reflexivity; and (5) the emotional labour of working alongside AI in crisis contexts. Participants consistently described AI as helpful for structure but limited in its support for the trauma-informed, culturally responsive, and ethically attuned dimensions of crisis care that exceed the capacity of current systems to process. CONCLUSIONS: The study concludes that AI cannot replace the embodied, intuitive, and relational dimensions of social work in crisis settings. Implications for practice include the need for human-centred AI design, transparent decision-support systems, and frameworks that preserve professional discretion. A relational and reflexive approach to AI integration is essential to ensure that technological innovation enhances rather than undermines the dignity and care at the heart of crisis social work.