Leveraging AI for Social Work Research: Possibilities for AI-Assisted Qualitative Data Analysis

Abstract

Can social workers harness the power of generative artificial intelligence (Al) as a tool for qualitative research? Might prompt engineering be a valuable skill of future social work researchers? This poster showcases how one researcher learned to overcome Al paralysis and embrace ChatPDF - a freely available Al application that uses natural language processing {NLP) to respond to user queries about an uploaded PDF. Examples experimenting with the Al tool ChatPDF for qualitative data analysis are presented, highlighting the advantages, pitfalls, and application-related considerations. As a two-phase curiosity experiment, the researcher engaged in a theory-building exercise to explore key concepts for understanding when using ChatPDF to assist researchers in qualitative data analysis, specifically in the organization of data into vignettes. The experiment generated insights about the purposeful use of Al tools for analysis and the risks of inaccuracy when researchers have not reached data familiarization or lack skills in prompting. The poster shares an assessment on whether ChatPDF and other NLPs are viable research assistants, the ethical issues with specific forms of qualitative data, and the potential of Al tools for social work students and community social work researchers.