Session: Standing at the Intersection of Social Work and Generative AI: Perspectives, Possibilities and Practices (Society for Social Work and Research 30th Annual Conference Anniversary)

This symposium presents four studies focusing on the integration of generative AI into social work practice. Each highlights the tensions between AI’s promise for the field and its potential risks. In these studies, clinicians report benefits including reduced administrative burden, enhanced clinical documentation, and novel insight comparable to supervisory feedback. Yet these potentials are tempered by concerns around confidentiality, data stewardship, and the erosion of human-centered care. Across qualitative interviews, platform evaluations, and social media discourse, practitioners express ambivalence and uncertainty, citing the lack of ethical guidelines, training opportunities, or regulatory guardrails.

This symposium questions whether AI will serve as a tool to help social work flourish through responsible integration, or contribute to its marginalization and eventual obsolescence. While some clinicians embrace AI as a supportive tool, others resist its presence entirely, warning of its dehumanizing potential. These findings highlight a crossroads for the profession: one calling for transformative social work leadership to ensure AI advances human dignity, equity, and ethical care in an evolving clinical landscape.

Study One explores how social workers are experiencing the infusion of AI into practice. Through interviews with 21 practitioners, researchers found both enthusiasm for AI’s potential to aid clinical tasks and concern about confidentiality, and professional identity. A key tension emerged: whether AI will supplant core social work functions or be ethically integrated to enhance practice. Findings highlight the need for human-centered approaches and professional leadership in shaping AI’s future.

Study Two explored therapists’ experiences using an AI platform to generate clinical notes. Interviews with 17 clinicians revealed benefits like reduced documentation burden, improved note quality , and new opportunities for reflective practice. However, participants raised concerns about data security, confidentiality, and the lack of ethical guidelines. While AI notes were deemed generally accurate, missing policies and practices hindered wider adoption. The findings highlight both the promise and the regulatory gaps surrounding AI use in behavioral health.

Study Three demonstrates fine-tuning of large language models (LLMs) for mental health support. While fine-tuning of models show strong readability, coherence, and empathy, there is a need for careful evaluation, ongoing refinement, and ethical oversight to ensure these tools support effective, responsible integration into mental health practice.

Study Four qualitatively analyzed 6,000 Reddit posts exploring therapists’ perspectives on generative AI in mental health. Findings revealed polarized views: some embraced AI for administrative and clinical support, while others rejected its role. Themes included ethical concerns, emotional reactions, societal impacts, and AI’s potential and limitations. Despite limited training, many clinicians are already using AI tools without clear guidance, highlighting the need for professional support frameworks.

Discussant: Bryan Victor, Ph.D. is an associate professor in the School of Social Work at Wayne State University. His recent work explores the integration of generative AI tools into social work practice and research.