Press freedom meets the machine on deadline
World Press Freedom Day examined artificial intelligence in journalism, just after the software suggested renaming the event “Seven Shocking Freedoms Editors Cannot Believe”.
UNESCO’s programme covered useful tools for investigation, translation and fact-checking. It also addressed deepfakes, surveillance, misinformation and the possibility that a machine may consume journalism without paying for lunch.
Independent reporting needs safety, trust and a viable business. Meanwhile, the newsroom chatbot needs sixteen seconds, three disclaimers and another reporter’s entire archive.
“AI can summarise a council meeting, but it cannot smell when the agenda is hiding something,” said Constance Byline, Editor of The Suspiciously Quiet Gazette. “Nor will it attend planning committee in a leisure centre on wet Tuesday.”
The debate matters because technology can widen access while concentrating power. A deepfake can travel globally before the correction has found its glasses and logged into the content system.
At the Bureau of Human Curiosity, reporters tested a new model on the question “Who benefits?” It replied that the request exceeded its context window and offered a recipe for soup.
Journalists will keep using new tools. They will also keep making calls, checking documents and asking unwelcome questions. The machine received a right of reply and generated four paragraphs beginning “In today’s rapidly evolving landscape”.
Read the real story: UNESCO’s World Press Freedom Day programme on AI and journalism.
The useful question is not whether software can write, but whether it can be bothered to ring a reluctant official for the third time. That remains journalism’s unlikely advantage. A machine can produce infinite copy about a scandal. It has yet to spend an afternoon outside a town hall discovering the press office has gone home early.
