Table of Contents

LEGAL

AI-Generated Content Disclosure Policy

v1Effective Date 2026-08-27

1. Mandatory Disclosure

Every asset created and traded on Castera carries an AI-generated label. Article 31 of Korea's Framework Act on the Development of Artificial Intelligence and Establishment of a Foundation for Trust (in force since 22 January 2026) and Article 23 of its Enforcement Decree (in force since 21 July 2026) govern the labeling of generative-AI output; the obligated party is the AI business operator providing the AI product or service. Castera treats this label as a precondition of every deal and guides creators and advertisers accordingly. This policy is not legal advice.

2. Disclosure Standards

Labels use at least one of two methods. (1) Human-perceivable — a corner mark or badge on images, a caption from the first frame of videos, an 'AI-generated' statement in the post text. (2) Machine-readable — file metadata. When only the machine-readable method is used, the fact of generation must also be communicated at least once in a notice (Enforcement Decree Art. 23(2)). Where a realistic human figure appears, a human-perceivable label is used (Act Art. 31(3)). Platform-native AI labeling (Instagram, TikTok, YouTube, etc.) is used alongside.

3. Brand Advertiser Responsibility

Advertisers do not remove the AI label from delivered assets when running ads (technical edits such as resizing, cropping, and compression are allowed). Responsibility for the truthfulness of the ad, efficacy claims, and consumer-protection law rests with the advertiser who determined the ad content; Castera provides the labeling criteria and does not accept delivery of unlabeled assets.

4. Enforcement

Violations of this policy may result in warnings, service restrictions, or account suspension at our discretion.

This document is currently under legal review, and some content may be adjusted based on the review outcome.