"We Don't Want to Remove Its Humanness": Aligning AI Large Language Model Use with NASW Core Values

Introduction: AI large language model (AI LLM) offer clear benefits in social work practice, including reducing time spent on repetitive administrative tasks, improving email communication, and providing concise, accessible summaries of complex policies, regulations, and research articles. However, social work is a human-centered profession that has historically served clients who have experienced marginalization, including loss of rights, poverty, and mental health challenges. Many practitioners feel threatened by AI LLM, which may dehumanize client-clinician relationships and create tensions with core professional values. This study examines how social workers can navigate the use of AI LLM to improve efficiency while upholding the six core values of the National Association of Social Workers (NASW): service, social justice, dignity and worth of the person, importance of human relationships, integrity, and competence.

Methods: Participants included nine students (4 BSW and 5 MSW), eleven faculty and staff (1 tenure-track, 7 non-tenure-track, 3 staff members), and eight practicum supervisors (4 behavioral health, 3 school-based, and 1 policy setting). Participants took part in one of two 90-minute Zoom focus groups. Faculty and staff reported varied tenure (55% with ≤9 years; 45% with ≥10 years), while practicum supervisors reported extensive experience (38% with 10–19 years; 63% with ≥20 years). Participants were recruited from a large school of social work offering online and in-person programs, with approximately 2,700 students enrolled in the 2025–2026 academic year. During the sessions, participants responded to prompts displayed on slides addressing how AI LLM could support or challenge each of the six NASW core values. Each participant was given 30 seconds to respond, followed by a 3-minute group discussion. Thematic analysis was used to identify recurring themes.

Results: Participants identified both opportunities and concerns related to AI LLM use across the six core values. For service, AI was seen as enhancing access to information, supporting data analysis and interpretation, reducing documentation burden, and improving care coordination. For social justice, participants noted the potential to reduce language barriers through translation tools, though concerns were raised about cultural responsiveness being undermined. Regarding dignity and worth of the person and the importance of human relationships, participants expressed concerns about confidentiality and data security, and the potential loss of human connection between a clinician and a client. To uphold integrity, participants emphasized the need for transparency and informed consent when using AI tools in practice. For competence, AI was viewed as a valuable tool for continuing education and knowledge access. However, participants also warned that reliance on AI without proper training or supervision could create a false sense of expertise.

Discussion and Conclusions: The NASW core values provide a critical ethical framework for navigating the integration of AI LLM in social work practice. However, participants demonstrated considerable variability in how these values are applied, often shaped by practice context and professional experience. Findings highlight the need for structured training and pedagogical approaches to support ethical, responsible, and value-consistent use of AI in social work.