[Tech World News - Reporter Myeong-e Jo] Twigfarm announced that it will open an API-based personal data anonymization service for large-scale data.

Companies that intend to incorporate artificial intelligence based on data containing personal information must first anonymize the personal data they possess before applying it to research or services. This is because the Data Act 3 has been revised and various regulations, notices, and guidelines related to anonymization have been released.

Anonymization topic that is not well addressed in artificial intelligence research institutions. There is a shortage of professionals specializing in anonymization in the artificial intelligence industry, and the demand for such expertise is higher than the supply. As a result, companies face difficulties in establishing dedicated teams for personal data anonymization. Without a dedicated team, companies find it challenging to enhance their understanding of relevant laws and regulations and lack the capacity to evaluate and adopt anonymization solutions.

CEO Sun-ho Baek stated, "Existing personal data anonymization solutions are provided as package software, which lacks scalability when dealing with large-scale data. On the other hand, API-based solutions with good usability and scalability from overseas often have low quality in terms of anonymization for the Korean language, making it difficult for companies to apply such solutions." The anonymization service released by Twigfarm is provided through an API-based platform, making it easier to anonymize large-scale data.

While participating in the AI Learning Data Construction Project hosted by the National IT Industry Promotion Agency (NIA), Twigfarm has accumulated expertise in personal data anonymization technology. For structured data with a consistent structure, anonymization based on rules is possible and provides excellent quality. However, there are technical limitations when it comes to anonymizing unstructured data based on rules. Twigfarm has improved the quality of anonymization for unstructured data in the Korean language by applying its self-developed Named Entity Recognition (NER) model.

Source: Tech World News (http://www.epnc.co.kr)

‍