DeepBrain AI Unveils Invisible Watermark Solution for 'AI Basic Law' Compliance

DeepBrainAI, a company specializing in generative AI, has launched a content identification solution in line with the implementation of the AI Basic Act. To comply with the AI Basic Act, which comes into effect on the 22nd, DeepBrainAI has released the 'Invisible Watermark' solution. This solution launch is designed to technically support the 'obligation to label AI-generated content' stipulated by the AI Basic Act. According to this legislation, all content created through generative AI, such as text, voice, video, and images, must include measures to identify whether it was generated by AI. Violation of this rule could result in a fine of up to 30 million won. The invisible watermark technology disclosed by DeepBrainAI inserts information into AI-generated content that is not visually identifiable. It leaves data regarding creation status and source within the file, allowing for tracking even after the content has been edited, copied, or recorded. Currently, this technology is applied in DeepBrainAI's generative AI platform, 'AI Studio,' and the company plans to expand its offerings to include an on-premise setup for businesses to build on their own servers in the future. In addition, DeepBrainAI provides features such as the insertion of AI-generated metadata (model name, creation time, etc.), prior notification and output labeling, and a five-year retention function for generation and risk records to ensure compliance with regulations. Furthermore, the company also supports a post-verification system through its deepfake detection solution, 'AI Detector.' CEO Jang Se-young of DeepBrainAI stated, "With the implementation of the AI Basic Act, the obligation for companies to manage the sources of their content and ensure transparency has been strengthened, and we will provide solutions so that businesses can technically respond to changes in the regulatory environment."