Komdigi Studies a Ministerial Rule Requiring AI Developers to Label Synthetic Content
Indonesia’s Ministry of Communication and Digital Affairs is studying a ministerial regulation that would require companies developing artificial intelligence to label the content their systems generate, Deputy Minister Nezar Patria said on Saturday, October 17. Speaking at the Indonesian Public Relations Convention 2026 in Surakarta, he said the line between reality and synthetic material keeps getting thinner as AI-made video, images and voices become harder to tell apart from the real thing, according to a ministry press release.
Nezar pointed to a fraud in Hong Kong about two years ago, in which a finance employee made 15 transfers adding up to 25 million US dollars after a video meeting in which every other participant, including the company’s senior leaders, was a digital fake. Cases like that, he said, show how realistic synthetic media can be turned to scams and information manipulation.
The approach under study borrows from authentication and provenance work by global technology companies, which embed metadata so that information about a file’s origin survives when an image or clip is processed with AI tools. Under the planned rule, developers would be expected to attach markers that separate engineered content from authentic material. The ministry has not yet published a draft text or a timetable.
What it means in practice
Labelling would sit alongside the wider governance plan in Komdigi’s adaptive national AI roadmap to 2029, which Nezar said is being drafted with more than 400 institutions, communities, ministries and industry players for the 2026 to 2029 period. It also extends the ministry’s stated preference for safety by design and sector-by-sector rules over slowing AI down, while complaints about misuse could flow to bodies such as PANDI’s planned AI-SIRT.
For newsrooms, platforms and public-relations teams, the practical question is who carries the burden. A developer-side duty puts the first obligation on model makers, but labels can be stripped when content is cropped or re-uploaded, and smaller local developers may struggle with technical standards that larger firms are still building. Nezar also told the convention that AI should be treated as a working partner rather than a substitute for human reasoning and accountability, and that credibility becomes the scarcest asset when speech can be faked by machines.
What to watch: the scope of any draft regulation, whether it covers foreign platforms as well as domestic developers, and whether it names a technical standard for provenance metadata.