AI transparency in the news: a simple label is not enough
New research shows that readers require clear, visible, and detailed disclosures when AI is used in journalism. A generic “AI-generated” label is not sufficient.
The public wants to know exactly how the technology was used: was it light copy-editing, or was the entire article or accompanying image created by AI? Visible logos, watermarks, and prominent placement make transparency much more meaningful.
However, a so-called ‘transparency dilemma’ arises. While openness increases awareness, it can simultaneously provoke skepticism and reduce trust in the news.
The key lesson for Responsible AI is therefore that transparency must provide context, rather than merely being a compliance checkmark. Readers must understand the role AI played, how human oversight was managed, and how reliable the information is.