Leveraging Retrieval-Augmented Generation (RAG) Technology for Enhanced Accreditation Preparedness and Compliance in Physical Therapy Education

BACKGROUND AND PURPOSE: Accreditation ensures educational program quality, public accountability, and professional readiness. Compliance demands substantial personnel effort and coordination. Although advances in artificial intelligence (AI), and more specifically large language models (LLMs), could streamline accreditation preparation, these models pose risks by producing plausible but inaccurate outputs, known as hallucinations. To mitigate risks, Retrieval-Augmented Generation (RAG) can improve the reliability and factual grounding of LLMs by incorporating retrieved source documents. Studies have shown that LLMs equipped with a RAG system reduce hallucinations and increase response accuracy. POSITION AND RATIONALE: The aim of this position paper is to address the challenge of resource-intensive and human error-prone accreditation preparation by advocating for the strategic integration of RAG technology. A RAG system enables programs to analyze internal documents against accreditation standards to generate curricular maps, reports, and recommendations. Effective use depends upon the quality of source documents and prompts. Adoption of a RAG system requires examining university policies and program-specific needs and ensuring human oversight. DISCUSSION AND CONCLUSION: The authors advocate for strategic integration of RAG technology for accreditation preparedness and continuous quality improvement in PT education. Institutions and stakeholders are encouraged to explore, implement, and evaluate RAG solutions to enhance operational efficiency, ensure data accuracy, and increase responsiveness to evolving educational and regulatory standards.