SelectStar said on Thursday it signed an MOU with the Korea Publication Culture Association to build a system to supply publishing data for AI training.
The two sides will 추진: refining, processing and structuring publishing data and verifying quality; setting up a data supply system for AI training; developing demand for publishing data and activating transactions; and jointly participating in AI data projects by the government and public institutions.
The association will consult with publishers and serve as the channel for collecting, managing and supplying data. SelectStar will use optical character recognition (OCR) and document layout analysis (DLA) technologies to support refining and structuring publishing data and verifying quality.
Data use will be based on consent from copyright holders and publishing rights holders. Actual supply and usage conditions will be set through separate contracts, and security measures such as access controls and encryption are also planned.
SelectStar plans to apply document preprocessing technologies accumulated through its existing AI training data projects to this cooperation. The technologies analyze complex document structures and process them into data suitable for AI training, and can handle documents in closed networks or on-premise environments without exporting them outside.
SelectStar CEO Kim Se-youp (김세엽) said, "Publishing content is a knowledge resource with high value for AI training, but it must be used based on the consent and trust of rights holders." He added, "Through experience building AI training data and document preprocessing technologies, we will lay the groundwork for the publishing industry and the AI industry to grow together."