Most news organisations treat AI as an engineering project with editorial consultation. That’s backwards. From input quality and metadata governance to retrieval logic and output tone, every stage of AI implementation demands editorial judgment — not as an afterthought, but as the driving force. Four specific responsibilities show how editorial teams must lead.
AI companions don’t display links — they synthesise answers. That shift from retrieval to generation means journalism now competes on machine-readability, not just editorial quality. Language Model Optimisation (LMO) is the discipline that closes the gap: a dual-text approach that keeps published journalism human-first while creating machine-optimised versions that power your own AI products.
People are replacing human advisors with AI companions for everything from career guidance to camera repair. News organisations have an underappreciated edge: audiences already consume their expertise without human contact, and their verified, editorially curated datasets offer something general-purpose AI cannot. The strategic shift is from reporting what happened to guiding what to do about it.