Revolutionizing Transportation: Kakao Mobility’s Level 4 Autonomous Driving Strategy with Physical AI

Kakao Mobility is embarking on a journey to develop Level 4 autonomous driving technologies internally as part of its physical AI strategy. The Vice President of Kakao Mobility’s Physical AI division, Kim Jin-kyu, unveiled the roadmap at the 2026 World IT Show conference in Seoul’s COEX. The focus of the presentation was on autonomous driving services integrated into mobility platforms in the physical AI era.

The event, themed “Beyond Idea, Into Action: AI moves Reality,” saw participation from 460 companies and organizations from 17 countries, according to Yonhap. South Korea’s Ministry of Science and ICT highlighted the event’s connection to a broader physical AI transition, where AI is applied in physical industrial sectors.

Kakao Mobility aims to merge autonomous driving technologies with physical infrastructure as part of its mobility strategy in Korea. The company is working towards establishing an open autonomous driving ecosystem to enhance local competitiveness.

In the realm of Level 4 autonomy, systems are capable of managing driving in specified service areas without the need for passenger monitoring or intervention, as defined by the US National Highway Traffic Safety Administration. These systems are typically deployed in designated service zones like autonomous taxi areas or fixed districts.

Kakao Mobility’s Level 4 roadmap is structured around three key technology areas: machine learning models, vehicle redundancy, and validation systems. The company is developing machine learning models to handle perception, decision-making, and control autonomously without human input. These functions encompass how autonomous vehicles perceive their surroundings, make driving decisions, and control movement.

Additionally, Kakao Mobility plans to implement vehicle architectures with redundant systems to ensure core functions can continue operating in the event of a critical component failure. The company’s validation platform will merge virtual simulations with real-world driving data to facilitate testing, performance enhancement, and quality assurance while developing autonomous driving services.

To enhance safety and control, Kakao Mobility is constructing an integrated safety management platform for autonomous vehicles. This includes the Autonomous Vehicle Visualizer, a 3D visualization tool enabling real-time sharing of a vehicle’s field of view to allow passengers to monitor driving conditions.

Moreover, Kakao Mobility intends to establish a 24-hour control center and an anomaly detection system using vision-language models to support real-time context analysis, remote intervention, and emergency response. The anomaly detection system will utilize vision-language models for monitoring autonomous driving services post-deployment.

In a bid to foster an open ecosystem, Kakao Mobility plans to share selected technology assets with companies, startups, and manufacturers working on autonomous driving. These assets encompass large-scale autonomous driving datasets, high-definition maps, and platform APIs for ride-hailing and dispatch services.

The company’s asset-sharing initiative aims to enable industry participants to develop autonomous driving technologies without independently building the underlying infrastructure. Kakao Mobility also intends to share operational resources such as fleet management systems and on-site response capabilities to support an open domestic autonomous driving ecosystem.

Highlighting their current efforts, Kakao Mobility referenced their late-night autonomous vehicle service in Seoul’s Gangnam district. The service, accessible through the Kakao T platform, offers users autonomous driving services alongside existing mobility options.

Since its launch on September 26, 2024, the Gangnam late-night autonomous taxi service recorded 7,754 rides until February 28, 2026, with no accidents attributed to autonomous driving technology during that period. The service transitioned from a free pilot to a paid operation in April 2026, expanding the fleet from three vehicles to seven, excluding two reserve vehicles.

Users can access the service through the Seoul Autonomous Car icon or the regular taxi-hailing menu on Kakao T, which consolidates multiple mobility services including taxi, navigation, and vehicle-related services. The Gangnam service is integrated into Kakao Mobility’s existing mobility platform.

In conclusion, Kakao Mobility’s foray into Level 4 autonomous driving technologies and the integration of physical AI in the mobility sector signifies a significant step towards establishing an open autonomous driving ecosystem. Through innovative technology development and strategic partnerships, the company aims to enhance safety, efficiency, and accessibility in autonomous driving services.