1. Home
  2. RIP
  3. View Record
https://www.nationalacademies.org/trb/events

A Generative AI Framework for Managing Public Comments in Transportation Agency Assessments

This research develops a Generative artificial intelligence (AI) framework for evaluating transportation agency resolution of public complaints and comments, addressing the challenge of siloed datasets where public feedback and agency improvement records remain disconnected and difficult to analyze collectively. Building on previous work demonstrating Large Language Model efficiency in analyzing public feedback, the study creates automated systems to identify patterns and correlations between reported concerns and documented improvements. The methodology involves collecting complementary datasets including public complaint narratives with location details and timestamps, alongside agency activity records documenting improvement efforts and outcomes. Natural Language Processing techniques will clean and standardize unstructured text data, while machine learning algorithms generate text embeddings and cluster recurring themes in complaints and agency responses. Large Language Models will perform semantic matching to quantify correlations between complaints and improvements, classify complaint-response pairs by resolution status, and conduct gap analysis identifying unaddressed service issues. Evaluation metrics include response time quantification, resolution effectiveness assessment, sentiment analysis of follow-up feedback, and identification of systemic gaps in agency responsiveness. The research produces a visual dashboard displaying complaint trends, response patterns, and automated reports providing actionable insights for transportation agencies and policymakers.