Key points are not available for this paper at this time.
Evidence synthesis methods, including meta-analysis and systematic reviews, are essential for advancing knowledge, informing practice, and guiding decisions in Human Resource Development (HRD). Given HRD’s reliance on non-experimental research, robust secondary data analysis is critical for enhancing evidence quality. Yet, manual data extraction remains a laborious, error-prone, barrier to scalability. An international panel of HRD experts (n = 26) engaged in iterative rounds of this mixed-method Delphi study exploring AI-supported data extraction in HRD research. Quantitative analysis followed Delphi protocols with pre-defined thresholds, and qualitative analysis included k-means clustering, thematic analysis, and sentiment analysis. The study pursued three objectives: (1) identify critical data elements and reporting structures for building a reliable HRD evidence base; (2) assess existing tools and define features of an ideal AI-assisted extraction tool; and (3) examine contextual factors that facilitate or hinder automation, ensuring applicability across diverse research settings. Findings revealed technological, disciplinary, and governance challenges. Experts prioritized methodological details (e.g., sample size, inclusion criteria) and expressed divergent views on tabular extraction. Concerns around bias, transparency, and ethics underscore the need for responsible AI stewardship. Key opportunities include bridging skill gaps, standardizing governance, and fostering interdisciplinary collaboration. HRD is uniquely positioned to shape ethical AI adoption and safeguard research quality.
Amanda Legate (Sun,) studied this question.