🚀 Level 4 Autonomy: Korea's Bold Shift 🇰🇷

KAKAO MOBILITY’S LEVEL 4 AUTONOMOUS DRIVING STRATEGY
Kakao Mobility is pursuing in-house development of Level 4 autonomous driving technologies as a core component of its broader physical AI strategy. Vice President and Head of Physical AI, Kim Jin-kyu, unveiled this roadmap at the 2026 World IT Show conference in Seoul, focusing on autonomous driving services integrated within mobility platforms for the “physical AI era.” The event, attended by 460 companies from 17 countries, underscored South Korea’s Ministry of Science and ICT’s commitment to a wider physical AI transition, highlighting Kakao Mobility’s ambition to establish an open autonomous driving ecosystem to bolster domestic competitiveness.

LEVEL 4 AUTONOMATION FRAMEWORK
Level 4 autonomy, according to the US National Highway Traffic Safety Administration, allows vehicles to operate without requiring passenger intervention in defined service areas. These systems typically function within limited zones, such as autonomous taxi zones or specific districts. Kakao Mobility’s Level 4 roadmap centers around three key technological pillars: machine learning models, vehicle redundancy systems, and robust validation platforms. The company’s primary focus is on developing machine learning models capable of handling perception, decision-making, and control autonomously, effectively enabling the vehicle to interpret its surroundings, make driving decisions, and manage movement without human input.

VEHICLE REDUNDANCY AND VALIDATION SYSTEMS
To ensure operational safety and reliability, Kakao Mobility plans to incorporate vehicle architectures with redundant systems. This design allows critical functions to continue operating even if a primary component fails, enhancing system resilience. The company’s validation platform will leverage a combined approach of virtual simulations and real-world driving data. This system is designed to facilitate rigorous testing, performance evaluation, and continuous improvement of the autonomous driving technology.

INTEGRATED SAFETY MANAGEMENT PLATFORM
Kakao Mobility is developing an integrated safety management platform specifically tailored for autonomous vehicles. A central element of this platform is the Autonomous Vehicle Visualizer, a 3D visualization tool that provides passengers with a real-time view of the vehicle’s surroundings. This tool displays the vehicle's field of view, allowing passengers to monitor driving conditions and understand the vehicle’s operational context during a ride.

REAL-TIME CONTEXT ANALYSIS AND REMOTE INTERVENTION
Looking ahead, Kakao Mobility plans to integrate a 24-hour control center equipped with anomaly detection systems utilizing vision-language models. These systems are intended to perform real-time context analysis, facilitate remote intervention, and enable rapid emergency response. The control center will continuously monitor autonomous driving services post-deployment, providing an additional layer of safety and operational oversight.

VISION-LANGUAGE MODEL DEVELOPMENT
The anomaly detection system will employ vision-language models; however, specific details regarding model architecture and performance metrics remain undisclosed at this time. The company’s intention is to leverage the power of AI to proactively identify and address potential issues, ensuring seamless and safe operation of the autonomous vehicles.

OPEN ECOSYSTEM STRATEGY
Kakao Mobility intends to foster collaboration within the autonomous driving sector by sharing selected technology assets with companies, startups, and manufacturers. These assets include large-scale autonomous driving datasets, high-definition (HD) maps, and platform APIs for ride-hailing and dispatch services. HD maps, crucial for autonomous driving, provide detailed road information used for localization and driving decisions.

RESOURCE SHARING AND COLLABORATION
Beyond technology assets, Kakao Mobility plans to share operational resources, including fleet management systems and on-site response capabilities, contributing to a comprehensive open domestic autonomous driving ecosystem. This collaborative approach aims to accelerate the development and deployment of autonomous driving solutions within South Korea.

GANGNAM SERVICE DATA AND PERFORMANCE
Kakao Mobility’s late-night autonomous vehicle service in Seoul’s Gangnam district serves as a key demonstration of its current capabilities. Accessible through the Kakao T platform, the service allows users to access autonomous driving services alongside existing mobility options. As of February 28, 2026, the Gangnam late-night autonomous taxi service had recorded 7,754 rides since its launch on September 26, 2024.

SERVICE EXPANSION AND ACCIDENT-FREE OPERATION
During this period, the Seoul Metropolitan Government reported zero accidents attributed to the autonomous driving technology. The service averaged approximately 24 trips per operating day. The service transitioned from a free pilot to a paid operation in April 2026, and the city expanded the fleet from three to seven vehicles, including two reserve vehicles. Users can access the service through Kakao T, utilizing either the Seoul Autonomous Car icon or the standard taxi-hailing menu, which integrates multiple mobility services within a single app.