Nota, an AI optimization technology company, is participating in a government on-device AI urban safety network demonstration project worth approximately 10.7 billion won. Moving away from the existing disaster management system that responded after analyzing data collected by a central control center, this initiative establishes a system where AI directly recognizes dangerous situations on-site and responds immediately, which is expected to expand the scope of AI utilization in the public safety sector.

Nota announced on the 8th that it has been selected as a participating company for the 'Demonstration and Expansion of Environment-Adaptive Multipurpose On-Device AI-Based Urban Safety Network' project under the '2026 On-Device AI Service Demonstration and Expansion Project' promoted by the Ministry of Science and ICT and the National Information & Communication Technology Promotion Agency (NIPA). This project is led by Chungnam Technopark and is being implemented by a consortium formed by Chungcheongnam-do, Cheonan City, and others.

The project sites are areas within Cheonan City that require disaster and public safety response, such as rivers, underpasses, riverside roads, and high-crime areas. The goal is to detect dangerous situations, such as flooding, vehicle entry, and abnormal behavior, at an early stage by utilizing sensor and CCTV data.

This project is designed to allow AI terminals to assess situations on-site, replacing the existing method of sending data to a central server for analysis. When a dangerous situation is detected, it can respond immediately by linking with on-site equipment such as circuit breakers, electronic display boards, and speakers, thereby contributing to reducing initial response time in the event of a disaster.

In this project, Nota is responsible for developing a Vision Language Model (VLM)-based complex risk detection AI model and optimizing it for on-device environments. Utilizing its proprietary AI model lightweighting and optimization platform, 'Netspresso,' Nota plans to optimize the VLM for the Mobilint NPU environment and implement a solution capable of analyzing various risk factors even in environments with limited power and computational resources.

On-device AI can process sensitive data, such as CCTV footage, on-site, thereby reducing the impact of communication delays and network failures. In terms of personal information protection, the ability to lower the burden compared to central server transmission is cited as a key reason for its application in public safety infrastructure.

This demonstration marks an instance where on-device AI is applied beyond industrial sites to urban safety infrastructure. Depending on the project results, there is also the potential for this model to expand into addressing common disaster and safety issues faced by local governments, such as rivers, underpasses, and high-crime areas.