Fairy Devices Selected for NEDO's Post-5G Project 'GENIAC' to Develop a Data Infrastructure for AI-based Evaluation of Work Quality in Manufacturing
Fairy Devices, Inc. announced its selection for NEDO's 'GENIAC' R&D project, part of the Post-5G Information and Communication Systems initiative. The company will focus on patrol-type work in manufacturing sites, conducting R&D to structure qualitative data—how a worker performs a task—into a format that AI can learn from. This initiative aims to promote digital transformation (DX) in Japanese manufacturing and contribute to the foundation for next-generation AI by providing the results as a public dataset.
📋 Article Processing Timeline
- 📰 Published: May 21, 2026 at 18:00
- 🔍 Collected: May 21, 2026 at 09:31
- 🤖 AI Analyzed: May 21, 2026 at 09:45 (13 min after Collected)
Fairy Devices, Inc. has announced that its proposal was selected for the 'Research and Development Project for AI-Ready Manufacturing Data (GENIAC),' a program publicly offered by the New Energy and Industrial Technology Development Organization (NEDO) under its 'R&D Project for Enhancement of the Post-5G Information and Communication Systems Infrastructure.'
Currently, the utilization of real-world data held by companies is crucial for improving AI performance. However, data from manufacturing sites is often managed in silos, and preparing it into an 'AI-Ready' state—properly structured and semantically annotated for AI consumption—remains a significant challenge.
In this project, Fairy Devices will focus on the 'quality evaluation' (how work is performed) of patrol-type tasks in manufacturing environments. The company will conduct research and development with several demonstration partner companies to structure this data into a format that AI can learn from and evaluate.
■ Details of the Selected Project
Project Name: R&D Project for Enhancement of the Post-5G Information and Communication Systems Infrastructure / Research and Development Project for AI-Ready Manufacturing Data (GENIAC)
Selected Theme: Construction of a Data Infrastructure for Describing Implementation Quality in Patrol-Type Work
Project Outline: The project will define qualitative aspects of patrol work in manufacturing that have been difficult to quantify, such as 'completeness of confirmation procedures,' 'adherence to safety conditions,' and 'stagnation or re-confirmation.' By establishing a hierarchical schema and annotation guidelines, the project will develop a reproducible, structured data infrastructure. The long-term goal is to connect this to interventional AI and multimodal foundation models.
Through this project, Fairy Devices aims to process information not just about 'what' a worker is doing, but 'how' they performed the work, into an AI-comprehensible format and provide it as a public dataset. This will contribute to the digital transformation (DX) of Japan's manufacturing sites and the establishment of a foundation for next-generation AI utilization.
■ Related Achievement: GENIAC-PRIZE Winner
In connection with this project, the company's AI agent technology won both the top '1st Place' prize and a special 'AI Agent Award' at the 'GENIAC-PRIZE,' hosted by the Ministry of Economy, Trade and Industry (METI) and NEDO. (Details: https://fairydevices.jp/20260324)
■ About Fairy Devices, Inc.
Operating under the vision 'Empower AI to Empower Human,' Fairy Devices provides a platform that transforms real-world multimodal information—such as audio, video, and sensor data—into structured data that AI can recognize and analyze. The company conducts R&D for industrial AI by leveraging data acquired from the field.
Currently, the utilization of real-world data held by companies is crucial for improving AI performance. However, data from manufacturing sites is often managed in silos, and preparing it into an 'AI-Ready' state—properly structured and semantically annotated for AI consumption—remains a significant challenge.
In this project, Fairy Devices will focus on the 'quality evaluation' (how work is performed) of patrol-type tasks in manufacturing environments. The company will conduct research and development with several demonstration partner companies to structure this data into a format that AI can learn from and evaluate.
■ Details of the Selected Project
Project Name: R&D Project for Enhancement of the Post-5G Information and Communication Systems Infrastructure / Research and Development Project for AI-Ready Manufacturing Data (GENIAC)
Selected Theme: Construction of a Data Infrastructure for Describing Implementation Quality in Patrol-Type Work
Project Outline: The project will define qualitative aspects of patrol work in manufacturing that have been difficult to quantify, such as 'completeness of confirmation procedures,' 'adherence to safety conditions,' and 'stagnation or re-confirmation.' By establishing a hierarchical schema and annotation guidelines, the project will develop a reproducible, structured data infrastructure. The long-term goal is to connect this to interventional AI and multimodal foundation models.
Through this project, Fairy Devices aims to process information not just about 'what' a worker is doing, but 'how' they performed the work, into an AI-comprehensible format and provide it as a public dataset. This will contribute to the digital transformation (DX) of Japan's manufacturing sites and the establishment of a foundation for next-generation AI utilization.
■ Related Achievement: GENIAC-PRIZE Winner
In connection with this project, the company's AI agent technology won both the top '1st Place' prize and a special 'AI Agent Award' at the 'GENIAC-PRIZE,' hosted by the Ministry of Economy, Trade and Industry (METI) and NEDO. (Details: https://fairydevices.jp/20260324)
■ About Fairy Devices, Inc.
Operating under the vision 'Empower AI to Empower Human,' Fairy Devices provides a platform that transforms real-world multimodal information—such as audio, video, and sensor data—into structured data that AI can recognize and analyze. The company conducts R&D for industrial AI by leveraging data acquired from the field.