The recent passage of the EU AI Act has made global regulations a reality. In particular, the requirements for high-risk AI systems are stringent, and it seems difficult for Korean companies to avoid the impact when entering the European market.
The biggest issues are transparency and explainability. In high-risk areas (credit scoring, hiring, healthcare, etc.), reasons for AI decisions must be provided by mandate. Korea has yet to establish its own legislation, so companies exporting to Europe need to prepare.
Additionally, generative AI now faces a watermarking requirement. This will apply to chatbots and image generation services, potentially adding a technical burden. In the short term, costs may rise, but in the long term, improved trustworthiness is expected.