Artificial Intelligence in Transportation

Chen, Sikai; Li, Pei; Sheng, Zihao; Ma, Junyi; Cheng, Yang; Qin, Xiao; Li, Yang; Shi, Tom; Roberts, Joel · 2025 · ROSA P / Wisconsin. Dept. of Transportation. Library and Research Unit

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This study provides a strategic assessment of integrating Artificial Intelligence (AI) into the operations of the Wisconsin Department of Transportation (WisDOT). Motivated by the need to address long-standing transportation challenges such as congestion, safety, and operational inefficiencies, the research aims to identify key opportunities, challenges, and implementation pathways for AI. The study focuses on six major domains: asset management, safety, operations, digital twins, autonomous vehicles, and generative AI. It seeks to map current AI capabilities, understand stakeholder perceptions, and inform strategic planning through evidence-based insights into benefit-risk tradeoffs and priority investment areas. The researchers employed a mixed-methods approach involving a comprehensive literature review, a nationwide stakeholder survey, and expert interviews. The literature review categorized national and international efforts across the six AI domains to summarize technical advances and operational considerations. The survey was distributed to over 100 transportation professionals across public agencies, academia, and private industry in 17 states, collecting data on AI readiness, data quality, and perceived benefits. This was supplemented by in-depth follow-up interviews with experts from state Departments of Transportation and engineering firms. The analysis included AI maturity mapping to evaluate applications based on data quality, implementation timelines, risk-benefit balance, and trustworthiness. Key findings reveal substantial divergence in AI perceptions between agency and non-agency stakeholders; state agencies prioritize trustworthiness and practical implementation, while academic stakeholders focus on innovation and technical readiness. Among the six domains, asset management and traffic operations exhibit the highest perceived readiness and return on investment, making them near-term priorities. Conversely, digital twins and generative AI are recognized for long-term potential but face higher uncertainty. Critical barriers to implementation include data fragmentation, lack of standardized formats, insufficient workforce training, and limited inter-agency collaboration. Additionally, high-benefit applications like safety analytics are associated with higher perceived risks regarding public trust and ethical concerns. The study concludes with a phased AI implementation roadmap for WisDOT. Short-term actions (1–3 years) focus on foundational steps such as improving data infrastructure, deploying proven AI tools for asset monitoring, and building internal capacity through training. Medium-term goals (4–7 years) involve scaling successful use cases and expanding to complex systems like real-time operations management. Long-term vision (8+ years) targets advanced integration, including predictive digital twins and autonomous infrastructure readiness. The report recommends establishing robust data governance, prioritizing initial deployments in asset management, safety, and operations, implementing tiered AI training programs, developing clear governance policies, and pursuing diversified partnership strategies to leverage external expertise.

Key finding

Asset management and traffic operations are identified as the highest-priority near-term AI applications for transportation agencies, while data fragmentation and workforce readiness represent the most critical barriers to broader implementation.

Methodology

mixed_methods

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• generative ai voice assistants