AI adoption rises as US cities monitor aging roads

Cities and states across the US are adopting AI technologies to monitor road conditions and prioritize repairs as infrastructure ages.

Hawaii is distributing 1,000 AI-enabled dashcams to drivers, and San Jose is expanding its AI system after it identified potholes with 97% accuracy.

Texas is also using AI tools, including dashcams and cellphone data, to inspect road signs and analyze risky driving behavior.

Massachusetts-based Cambridge Mobile Telematics works with several states to detect hazardous driving patterns and address infrastructure problems.

Local governments are sharing data and best practices through the GovAI Coalition, launched publicly in 2024.

Officials say these AI projects help identify hazards faster and support efforts to improve road safety as traffic fatalities rise in some states.

🧠 Food for thought

Implications, context, and why it matters.

San Jose in California reports 97% accuracy for AI-detected potholes in its own testing yet the news article lacks third-party evaluations that measure crash reductions, faster repairs, or lower maintenance costs at scale.
A 2025 case study reported 99.9% accuracy and a 99 F1-score (a balance of precision and recall) for International Roughness Index (IRI) classification using Random Forests (an ensemble machine learning method) 1. Deployment still runs into trouble with adjacent condition categories 1 and with lighting or image variability documented in the literature 2.
The shift from manual to semi and fully automated pavement condition surveys has been tracked over the last two decades 2.
Without standardized performance metrics and independent audits across jurisdictions, separating real infrastructure gains from enthusiasm for AI contracts remains hard.

The GovAI Coalition (a group of local governments that share data and best practices) promotes open standards and open-source resources 3. Platform teams can build privacy-first systems that ingest dashcam plus telematics feeds (vehicle and smartphone sensor data) from multiple vendors. They also handle redaction, retention, plus cross-agency sharing.
State Departments of Transportation (DOTs) use pavement surveys for maintenance planning, budget forecasting, and performance modeling 2.
A unified platform that aligns with emerging coalition standards could cut duplicate work when agencies negotiate alone with vendors 3.
Geographic Information Systems (GIS) with machine learning (ML) worked for pavement management in case studies 1.