Government to Create 'Vault' for Manufacturing Data Management for M.AX

The government is establishing a secure infrastructure for storing and managing manufacturing data to facilitate the transition to manufacturing artificial intelligence (M.AX). A key component of this initiative is the creation of a 'manufacturing data library' and the establishment of a 'clean room' to prevent external data leaks.

On June 5, the Ministry of Trade, Industry and Energy held the '3rd M.AX Expert Conference' with industry, academia, and research experts to announce these plans.

Amid the global AI competition, the government is promoting the advancement of the manufacturing sector, where South Korea has a competitive edge. High-quality manufacturing data held by companies is seen as a strategic asset essential for the country’s goal of becoming a leader in M.AX. The government believes that creating an environment and infrastructure where both manufacturing and AI companies can thrive is crucial for the success of M.AX.

To this end, the Ministry is working with over 1,500 manufacturing and AI companies, as well as academic and research institutions, to build an ecosystem for securing and utilizing manufacturing data through the M.AX Alliance, which consists of 11 sectors.

However, companies are understandably concerned about the potential leakage of their manufacturing data, which includes intellectual property such as core technologies and production know-how. For companies to provide data for subsequent research and development projects, a secure data storage and management infrastructure must be established first.

The government plans to create a 'manufacturing data library' to systematically manage the data provided by companies. This library will house high-quality manufacturing data and will implement specific devices and procedures to prevent data leaks and ensure security. Data will only be accessible within the clean room, which is isolated from external access, and external transfers will be prohibited. Access to the data will also require a separate review process.

This approach stems from the understanding that the key to AI competitiveness lies not in simple algorithms, but in high-quality training data. Recent analyses in the generative AI market indicate that differences in model performance are largely determined by the quality and scale of data, intensifying the competition for industrial data acquisition.

As the establishment of the library will take time, the Ministry has been using the 'Manufacturing AI Solution Development Support Center' operated by the Korea Electronics Technology Institute since May as a temporary base to store data acquired from AI factory projects. By the end of the year, the Ministry plans to develop a prototype of a manufacturing AI foundation model using this data and will conduct field applications and performance validations with participating companies in the M.AX Alliance.

Kim Sung-yeol, Director of the Ministry's Industrial Growth Office, stated, "The core of our industrial competitiveness in the AI era lies in manufacturing data and AI models tailored to specific industries. To collect and utilize high-quality manufacturing data, we must create an environment where both data-holding companies and AI firms can benefit mutually."

* This article has been translated by AI.

Copyright ⓒ Aju Press All rights reserved.