Scatter Lab has established a safety organization directly under its CEO to strengthen user protection for its generative artificial intelligence (AI) service, Zeta. The company plans to apply its own moderation technology throughout the service, from content review to interactions with AI, and expand safety standards to include restrictions on youth usage time and payment limits.On November 17, Scatter Lab announced the creation of the 'Trust & Safety Team,' which will be responsible for content safety and user protection for Zeta. The company intends to allocate half of its key AI development personnel to the development of content and conversation moderation AI models. It aims to make user and youth protection a core management priority, integrating the establishment of content safety policies, the development of its own AI models, and content technology and review tasks.Initially, the company will enhance the review system for content creation and distribution. By the end of this year, Scatter Lab plans to advance its own AI model and plot moderation functions to clearly classify content usage ratings and establish a protection system suitable for user ages.The plot review will be conducted in three stages: the creator's production process, the release timing, and the exposure timing within the service. At each stage, the content's material and text/images will be reviewed to classify usage ratings. AI content policies and review standards applicable to major service countries such as South Korea, Japan, and the United States will also be established.To achieve this, the company will develop a specialized AI model for plot review using large language models (LLMs). Age verification procedures will also be strengthened. Based on verified user ages, accessible plots will be differentiated to provide age-appropriate content from the initial registration.Safety measures will also be enhanced during conversations between users and AI. The company will launch its own AI model to prevent inappropriate conversations and establish response criteria based on user age, implementing additional measures to ensure compliance with these standards.A feature will be introduced to periodically inform users that they are interacting with AI. For domestic youth users, a daily usage limit of three hours will be set, and a basic payment limit of 100,000 won per month will be applied within the year. Additionally, a feature will be implemented to provide appropriate guidance to users when sensitive content related to suicide or self-harm is detected, following established criteria.* This article has been translated by AI.
