The Reuters Institute's annual Digital News Report reveals that 40% of online news articles now involve artificial intelligence in some stage of their production — from research and fact-checking to writing and headline generation.
The finding marks a dramatic increase from 12% in 2024 and 25% in 2025. AI involvement ranges from minor (suggesting headlines, checking facts) to substantial (drafting entire articles that are then edited by humans).
Reader trust varies significantly by disclosure. When articles are labelled as 'AI-assisted,' 61% of readers say they trust the content. When the same articles lack AI disclosure, trust rises to 78% — suggesting that AI labelling itself reduces perceived credibility.
The ethical implications are significant. If readers trust AI-assisted content less when they know about the AI involvement, disclosure requirements could paradoxically undermine trust in journalism at a time when trust is already low.