# Mobilintec develops AI road safety management model > Source: > Published: 2026-06-22 07:44:11.769078+00:00 # Mobilintec develops AI road safety management model The Korea Times reports that **Mobilintec**, a mobility and transportation data startup based in the University of Seoul campus town program, will participate as a joint research institution in the "2026 AI City Innovative Technology Discovery Project" organized by the **Ministry of Land, Infrastructure and Transport**. The Korea Times reports Mobilintec intends to develop and demonstrate an **agentic AI**-based traffic collision prediction and response system for local roads shared by pedestrians, vehicles, personal mobility devices and delivery robots. The Korea Times quotes Hong Ji-yeon, CEO of Mobilintec and a professor at the University of Seoul: "This project is a meaningful initiative to validate a new safety management model that uses agentic AI technology to proactively predict and respond to traffic risks on local roads." Editorial analysis: Government-led pilots such as this create operational testbeds that produce real-world datasets and evaluation scenarios for agentic AI in mixed-user urban streets. ### What happened The Korea Times reports that **Mobilintec**, a mobility and transportation data startup supported by the **University of Seoul** campus town program, will participate as a joint research institution in the "2026 AI City Innovative Technology Discovery Project," organized by the **Ministry of Land, Infrastructure and Transport**. The Korea Times reports Mobilintec plans to develop and demonstrate an **agentic AI**-based system for real-time traffic collision prediction and response on local roads shared by pedestrians, vehicles, personal mobility devices and delivery robots. The Korea Times quotes Hong Ji-yeon, CEO of Mobilintec and a professor in transportation engineering at the University of Seoul: "This project is a meaningful initiative to validate a new safety management model that uses agentic AI technology to proactively predict and respond to traffic risks on local roads." ### Editorial analysis - technical context Agentic AI in traffic management implies systems that sense, reason, and trigger responses with some degree of autonomy; industry observers describe these systems as combining perception, prediction, and decision modules. Deploying agentic AI on local roads raises recurring engineering requirements: low-latency sensor fusion across cameras and connected-device telemetry, robust multi-agent trajectory prediction for mixed road users, and safe action-selection under uncertainty. For practitioners, evaluation will require labeled near-miss and behavioral datasets that reflect pedestrian and micromobility interactions rather than highway traffic patterns. ### Context and significance Government-funded "AI city" pilots have become a common pathway for startups to access municipal data, run live demonstrations, and validate safety claims in constrained environments. Such pilots can surface integration challenges around edge compute, communications reliability, and cross-vendor interoperability. For cities, demonstrations that shift focus from post-accident response to proactive risk prediction change requirements for sensor coverage, data-sharing agreements, and incident verification workflows. ### What to watch Observers should watch for published demonstration metrics, the types of sensors and data sources Mobilintec integrates, and any third-party safety audits or verification the project adopts. Also monitor whether results are shared with municipal planners and whether the pilot produces reusable datasets or benchmarks for collision prediction on local roads. ## Scoring Rationale This is a regionally scoped government pilot that tests agentic AI in mixed-road environments, relevant to practitioners working on perception, prediction, and edge deployment. The story is notable but not transformational for the wider AI industry. Practice interview problems based on real data 1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with. [Try 250 free problems](/problems)