According to a report by ad-hoc-news.de, Taiwan's Ministry of Digital Affairs (MODA) is urging major online platforms to introduce a three-stage labeling system for AI-generated content. According to the report, the proposal was presented on 25 October 2026.
The model
In the first stage, creators must confirm via a selection box when uploading whether AI was used. In the second stage, platforms should automatically add an AI notice if their systems detect corresponding technical signals or metadata without a label having been applied. The third stage provides reporting channels through which authorities and users can flag unlabeled AI content so platforms can review it afterwards.
Visibility of notices
According to the report, MODA also calls for more prominent notices. For YouTube, the ministry asks, for example, that AI notices appear directly under the video rather than only in the expandable description field.
State of implementation
According to the report, the initiative followed a year of discussion in which, according to department director Tsai Shou-chuan, the authority exchanged views with Meta, Google, the messaging service Line, and industry associations. According to the authority, Google is making faster progress than Meta, which is still carrying out technical adjustments. At Line, the obligation is for now to apply only to publicly accessible content, not private messages.
Context
If the model is adopted, it could become an example of how labeling obligations outside the EU are concretely shaped. For German companies that publish AI content on platforms, the question is whether similar requirements would follow domestically; the source provides no statements on this.
Sources
Editorially owned by Ideal Syka. Sources and method: Newsroom & method. Tips and corrections: ai@i6eal.de.
