Korean comic creators are urging UNESCO to back consent and transparency safeguards for AI training, arguing that permission to publish a work should not automatically cover training a model on it. A survey cited by their association found that about 90 per cent of 164 respondents did not know how to stop their work being used for AI training.
Watch Desk analysis
What happened
Kwon Hyuk-joo, president of the creators’ association, made the case at an international conference at UNESCO headquarters in Paris. He said creators should be able to decide separately whether their comics can be published, used to train AI models, or used in AI services such as chatbots. He also called for labels on AI-generated content and transparency about training data.
The association’s survey found that 81.1 per cent of respondents would take part in an opt-in system requiring prior consent. UNESCO said the conference discussions would inform its digital policy work under the 2005 Convention on the Protection and Promotion of the Diversity of Cultural Expressions. The Korea Times reported the remarks and survey results; the survey figures are those reported by the association.
Why it matters
The distinction Kwon draws is practical: allowing a comic to be read is not necessarily the same as allowing it to help train a model, or letting its characters appear in a chatbot. That is a sharper question than treating all reuse as one permission slip.
The survey also points to a gap between creators’ appetite for consent-based use and their reported ability to prevent training use. Its 164 respondents are not proof of what every creator wants, but the figures give the policy discussion something more concrete than a general appeal to fairness.
Our read
Consent is meaningful only if creators can understand what they are agreeing to and make choices that apply to distinct uses. The proposal is not a finished rule, and UNESCO’s next steps are not settled. Still, separating publication, training and chatbot use is a useful starting point, rather than asking creators to accept one sprawling yes-or-no button.
What to watch
- Whether UNESCO’s policy work carries these proposals into concrete guidance.
- Whether creator groups develop opt-in systems that distinguish publication, training and AI-service use.
- Whether future surveys show creators can identify and control uses of their work.
Discussion spark: Should permission to publish a comic also cover training AI on it, or should those uses always require separate consent?
Sources and evidence
- Source update (30 October 2026, 05:35 UTC)
- Source update (30 October 2026, 05:35 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.